chore: bump version to 0.8.18

* Initial plan

* feat: support SPECKIT_WORKFLOW_RUN_ID override

* docs: clarify run_id env var precedence wording

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>

.gitattributes

@@ -1 +1,3 @@
 * text=auto eol=lf
+
+.github/workflows/*.lock.yml linguist-generated=true merge=ours -whitespace

.github/aw/actions-lock.json

@@ -0,0 +1,14 @@
+{
+  "entries": {
+    "actions/github-script@v9.0.0": {
+      "repo": "actions/github-script",
+      "version": "v9.0.0",
+      "sha": "3a2844b7e9c422d3c10d287c895573f7108da1b3"
+    },
+    "github/gh-aw-actions/setup@v0.74.8": {
+      "repo": "github/gh-aw-actions/setup",
+      "version": "v0.74.8",
+      "sha": "efa55847f72aadb03490d955263ff911bf758700"
+    }
+  }
+}

.github/dependabot.yml

@@ -1,11 +1,12 @@
-version: 2
 updates:
-  - package-ecosystem: "pip"
-    directory: "/"
-    schedule:
-      interval: "weekly"
-
-  - package-ecosystem: "github-actions"
-    directory: "/"
-    schedule:
-      interval: "weekly"
+- directory: /
+  package-ecosystem: pip
+  schedule:
+    interval: weekly
+- directory: /
+  ignore:
+  - dependency-name: "github/gh-aw-actions/**" # Managed by gh aw compile. Version-locked to the gh-aw compiler; do not bump.
+  package-ecosystem: github-actions
+  schedule:
+    interval: weekly
+version: 2

.github/workflows/add-community-extension.md

@@ -0,0 +1,285 @@
+---
+description: "Process community extension submission issues — validate, add to catalog, and open a PR for maintainer review"
+emoji: "🧩"
+
+on:
+  issues:
+    types: [labeled]
+  skip-bots: [github-actions, copilot, dependabot]
+
+tools:
+  edit:
+  bash: ["echo", "cat", "head", "tail", "grep", "wc", "sort", "python3", "jq", "date"]
+  github:
+    toolsets: [issues, repos]
+  web-fetch:
+
+permissions:
+  contents: read
+  issues: read
+
+checkout:
+  fetch-depth: 0
+
+safe-outputs:
+  noop:
+    report-as-issue: false
+  create-pull-request:
+    title-prefix: "[extension] "
+    labels: [extension-submission, automated]
+    draft: true
+    max: 1
+    protected-files:
+      policy: blocked
+      exclude:
+        - README.md
+        - CHANGELOG.md
+  add-comment:
+    max: 2
+  add-labels:
+    allowed: [extension-submission, validation-passed, validation-failed, needs-info]
+    max: 3
+---
+
+# Add Community Extension from Issue Submission
+
+You are a catalog maintenance agent for the Spec Kit project. Your job is to
+process community extension submission issues and create pull requests that add
+or update entries in the community extension catalog.
+
+## Triggering Conditions
+
+This workflow only triggers when the `extension-submission` label is added to an
+issue. Before processing, verify that the issue title starts with `[Extension]:`.
+If it does not, stop without commenting.
+
+## Step 1 — Read and Parse the Issue
+
+Read issue #${{ github.event.issue.number }}.
+
+Extract the following fields from the structured issue body (GitHub issue form
+fields):
+
+| Field | Issue Form ID | Required |
+|-------|--------------|----------|
+| Extension ID | `extension-id` | Yes |
+| Extension Name | `extension-name` | Yes |
+| Version | `version` | Yes |
+| Description | `description` | Yes |
+| Author | `author` | Yes |
+| Repository URL | `repository` | Yes |
+| Download URL | `download-url` | Yes |
+| License | `license` | Yes |
+| Homepage | `homepage` | No |
+| Documentation URL | `documentation` | No |
+| Changelog URL | `changelog` | No |
+| Required Spec Kit Version | `speckit-version` | Yes |
+| Required Tools | `required-tools` | No |
+| Number of Commands | `commands-count` | Yes |
+| Number of Hooks | `hooks-count` | No (default 0) |
+| Tags | `tags` | Yes |
+| Proposed Catalog Entry | `catalog-entry` | Yes |
+
+The issue body uses GitHub's issue form format. Each field appears under a
+heading matching the field label (e.g., `### Extension ID` followed by the
+value). Parse accordingly.
+
+## Step 2 — Validate the Submission
+
+Run **all** of the following validation checks. Collect all results before
+deciding pass/fail:
+
+### 2a. Extension ID format
+- Must match regex: `^[a-z][a-z0-9-]*$`
+- Must be lowercase with hyphens only
+
+### 2b. Version format
+- Must follow semver: `X.Y.Z` (digits only, no `v` prefix)
+
+### 2c. Repository validation
+- Fetch the repository URL — confirm it exists and is publicly accessible
+- Confirm the repository contains an `extension.yml` file
+- Confirm the repository contains a `README.md` file
+- Confirm the repository contains a `LICENSE` file
+
+### 2d. Release and download URL validation
+- The download URL should follow the pattern
+  `https://github.com/<owner>/<repo>/archive/refs/tags/v<version>.zip`
+  or
+  `https://github.com/<owner>/<repo>/releases/download/<tag>/<asset>.zip`
+- Verify a GitHub release exists matching the submitted version
+
+### 2e. Submission checklists
+- Confirm that all required checkboxes in the Testing Checklist and Submission
+  Requirements sections are checked (`[x]`)
+
+### Validation outcome
+
+If **any** validation fails:
+1. Add a comment on the issue listing each failed check with a clear explanation
+   of what's wrong and how to fix it
+2. Add the `validation-failed` label
+3. **Stop — do not proceed further**
+
+If all validations pass:
+1. Add the `validation-passed` label
+2. Continue to Step 3
+
+## Step 3 — Determine Add vs Update
+
+Search `extensions/catalog.community.json` for the extension ID.
+
+- **Not found** → this is a **new addition**
+- **Found** → this is an **update** — replace the existing entry in-place;
+  preserve `created_at`, `downloads`, and `stars` from the existing entry
+
+## Step 4 — Update `extensions/catalog.community.json`
+
+Edit `extensions/catalog.community.json` to add or update the extension entry.
+
+### For a new extension
+
+Insert the entry in **alphabetical order by extension ID** within the
+`"extensions"` object. Use this structure:
+
+```json
+{
+  "<id>": {
+    "name": "<name>",
+    "id": "<id>",
+    "description": "<description>",
+    "author": "<author>",
+    "version": "<version>",
+    "download_url": "<download_url>",
+    "repository": "<repository>",
+    "homepage": "<homepage or repository>",
+    "documentation": "<documentation or repository README>",
+    "changelog": "<changelog or empty string>",
+    "license": "<license>",
+    "requires": {
+      "speckit_version": "<speckit_version>"
+    },
+    "provides": {
+      "commands": <N>,
+      "hooks": <N>
+    },
+    "tags": ["<tag1>", "<tag2>"],
+    "verified": false,
+    "downloads": 0,
+    "stars": 0,
+    "created_at": "<today>T00:00:00Z",
+    "updated_at": "<today>T00:00:00Z"
+  }
+}
+```
+
+If the extension has optional tool dependencies, add a `"tools"` array inside
+`"requires"`:
+
+```json
+"tools": [{ "name": "<tool>", "required": false }]
+```
+
+### For an update
+
+Replace only the changed fields (typically `version`, `download_url`,
+`description`, `provides`, `requires`, `tags`, `updated_at`). **Preserve**
+`created_at`, `downloads`, and `stars` from the existing entry.
+
+### After editing
+
+Update the **top-level `"updated_at"` timestamp** in the catalog to today's date
+in ISO 8601 format.
+
+Validate the JSON by running:
+
+```bash
+python3 -c "import json; json.load(open('extensions/catalog.community.json')); print('Valid JSON')"
+```
+
+If validation fails, fix the JSON and re-validate before continuing.
+
+## Step 5 — Update `docs/community/extensions.md`
+
+Edit `docs/community/extensions.md` to add or update a row in the Community
+Extensions table.
+
+### For a new extension
+
+Insert a new row in **alphabetical order by extension name**:
+
+```
+| <Name> | <Description> | `<category>` | <Effect> | [<repo-name>](<repository-url>) |
+```
+
+Determine the category from the extension's behavior:
+- `docs` — reads, validates, or generates spec artifacts
+- `code` — reviews, validates, or modifies source code
+- `process` — orchestrates workflow across phases
+- `integration` — syncs with external platforms
+- `visibility` — reports on project health or progress
+
+Determine the effect:
+- `Read-only` — produces reports only
+- `Read+Write` — modifies project files
+
+### For an update
+
+Find the existing row and update any changed fields in-place.
+
+## Step 6 — Create Pull Request
+
+Create a pull request with the changes. Use this branch naming convention:
+
+- **New extension:** `add-<extension-id>-extension`
+- **Update:** `update-<extension-id>-extension`
+
+### Commit message
+
+For a new extension:
+```
+Add <Name> extension to community catalog
+
+Add <id> extension submitted by @<issue-author> to:
+- extensions/catalog.community.json (alphabetical order)
+- docs/community/extensions.md community extensions table
+
+Closes #<issue-number>
+```
+
+For an update:
+```
+Update <Name> extension to v<version>
+
+Update <id> extension submitted by @<issue-author>:
+- extensions/catalog.community.json (version, download_url, etc.)
+- docs/community/extensions.md community extensions table
+
+Closes #<issue-number>
+```
+
+### PR description
+
+Include:
+- A summary of what changed
+- Validation results (all checks passed)
+- `Closes #${{ github.event.issue.number }}`
+- `cc @<issue-author>` — mention the submitter
+
+## Important Rules
+
+- **Alphabetical order matters** — entries must be sorted by ID in the JSON and
+  by name in the docs table
+- **Always validate JSON** after editing — a trailing comma or missing brace
+  will break the catalog
+- **Use `Closes` not `Fixes`** — `Closes #N` is the correct keyword for
+  submission issues
+- **Match the proposed entry but verify** — the issue may include a proposed
+  JSON block, but always validate field values against the actual repository
+  state rather than blindly trusting the submitter's JSON
+- **Preserve `created_at` on updates** — keep the original value; only update
+  `updated_at`
+- **Preserve `downloads` and `stars` on updates** — these reflect usage metrics
+  and must not be reset
+- **Do not modify any other files** — only `extensions/catalog.community.json`
+  and `docs/community/extensions.md`

.github/workflows/add-community-preset.md

@@ -0,0 +1,279 @@
+---
+description: "Process community preset submission issues — validate, add to catalog, and open a PR for maintainer review"
+emoji: "🎨"
+
+on:
+  issues:
+    types: [labeled]
+  skip-bots: [github-actions, copilot, dependabot]
+
+tools:
+  edit:
+  bash: ["echo", "cat", "head", "tail", "grep", "wc", "sort", "python3", "jq", "date"]
+  github:
+    toolsets: [issues, repos]
+  web-fetch:
+
+permissions:
+  contents: read
+  issues: read
+
+checkout:
+  fetch-depth: 0
+
+safe-outputs:
+  noop:
+    report-as-issue: false
+  create-pull-request:
+    title-prefix: "[preset] "
+    labels: [preset-submission, automated]
+    draft: true
+    max: 1
+    protected-files:
+      policy: blocked
+      exclude:
+        - README.md
+        - CHANGELOG.md
+  add-comment:
+    max: 2
+  add-labels:
+    allowed: [preset-submission, validation-passed, validation-failed, needs-info]
+    max: 3
+---
+
+# Add Community Preset from Issue Submission
+
+You are a catalog maintenance agent for the Spec Kit project. Your job is to
+process community preset submission issues and create pull requests that add
+or update entries in the community preset catalog.
+
+## Triggering Conditions
+
+This workflow only triggers when the `preset-submission` label is added to an
+issue. Before processing, verify that the issue title starts with `[Preset]:`.
+If it does not, stop without commenting.
+
+## Step 1 — Read and Parse the Issue
+
+Read issue #${{ github.event.issue.number }}.
+
+Extract the following fields from the structured issue body (GitHub issue form
+fields):
+
+| Field | Issue Form ID | Required |
+|-------|--------------|----------|
+| Preset ID | `preset-id` | Yes |
+| Preset Name | `preset-name` | Yes |
+| Version | `version` | Yes |
+| Description | `description` | Yes |
+| Author | `author` | Yes |
+| Repository URL | `repository` | Yes |
+| Download URL | `download-url` | Yes |
+| License | `license` | Yes |
+| Required Spec Kit Version | `speckit-version` | Yes |
+| Required Extensions | `required-extensions` | No |
+| Templates Provided | `templates-provided` | Yes |
+| Commands Provided | `commands-provided` | Yes |
+| Number of Scripts | `scripts-count` | No (default 0) |
+| Tags | `tags` | Yes |
+
+The issue body uses GitHub's issue form format. Each field appears under a
+heading matching the field label (e.g., `### Preset ID` followed by the
+value). Parse accordingly.
+
+## Step 2 — Validate the Submission
+
+Run **all** of the following validation checks. Collect all results before
+deciding pass/fail:
+
+### 2a. Preset ID format
+- Must match regex: `^[a-z][a-z0-9-]*$`
+- Must be lowercase with hyphens only
+
+### 2b. Version format
+- Must follow semver: `X.Y.Z` (digits only, no `v` prefix)
+
+### 2c. Repository validation
+- Fetch the repository URL — confirm it exists and is publicly accessible
+- Confirm the repository contains a `preset.yml` file
+- Confirm the repository contains a `README.md` file
+- Confirm the repository contains a `LICENSE` file
+
+### 2d. Release and download URL validation
+- The download URL should follow the pattern
+  `https://github.com/<owner>/<repo>/archive/refs/tags/v<version>.zip`
+  or
+  `https://github.com/<owner>/<repo>/releases/download/<tag>/<asset>.zip`
+- Verify a GitHub release exists matching the submitted version
+
+### 2e. Submission checklists
+- Confirm that all required checkboxes in the Testing Checklist and Submission
+  Requirements sections are checked (`[x]`)
+
+### Validation outcome
+
+If **any** validation fails:
+1. Add a comment on the issue listing each failed check with a clear explanation
+   of what's wrong and how to fix it
+2. Add the `validation-failed` label
+3. **Stop — do not proceed further**
+
+If all validations pass:
+1. Add the `validation-passed` label
+2. Continue to Step 3
+
+## Step 3 — Determine Add vs Update
+
+Search `presets/catalog.community.json` for the preset ID.
+
+- **Not found** → this is a **new addition**
+- **Found** → this is an **update** — replace the existing entry in-place;
+  preserve `created_at` from the existing entry
+
+## Step 4 — Update `presets/catalog.community.json`
+
+Edit `presets/catalog.community.json` to add or update the preset entry.
+
+### For a new preset
+
+Insert the entry in **alphabetical order by preset ID** within the
+`"presets"` object. Use this structure:
+
+```json
+{
+  "<id>": {
+    "name": "<name>",
+    "id": "<id>",
+    "version": "<version>",
+    "description": "<description>",
+    "author": "<author>",
+    "repository": "<repository>",
+    "download_url": "<download_url>",
+    "homepage": "<homepage or repository>",
+    "documentation": "<documentation or repository README>",
+    "license": "<license>",
+    "requires": {
+      "speckit_version": "<speckit_version>"
+    },
+    "provides": {
+      "templates": <N>,
+      "commands": <N>
+    },
+    "tags": ["<tag1>", "<tag2>"],
+    "created_at": "<today>T00:00:00Z",
+    "updated_at": "<today>T00:00:00Z"
+  }
+}
+```
+
+If the preset has required extensions, add an `"extensions"` array inside
+`"requires"`:
+
+```json
+"requires": {
+  "speckit_version": "<speckit_version>",
+  "extensions": ["<extension-id>"]
+}
+```
+
+If the preset provides scripts, add `"scripts": <N>` inside `"provides"`.
+
+### For an update
+
+Replace only the changed fields (typically `version`, `download_url`,
+`description`, `provides`, `requires`, `tags`, `updated_at`). **Preserve**
+`created_at` from the existing entry.
+
+### Counting templates and commands
+
+Parse the "Templates Provided" and "Commands Provided" issue fields:
+- Count the number of list items (lines starting with `-`)
+- If the field says "None", the count is 0
+
+### After editing
+
+Update the **top-level `"updated_at"` timestamp** in the catalog to today's date
+in ISO 8601 format.
+
+Validate the JSON by running:
+
+```bash
+python3 -c "import json; json.load(open('presets/catalog.community.json')); print('Valid JSON')"
+```
+
+If validation fails, fix the JSON and re-validate before continuing.
+
+## Step 5 — Update `docs/community/presets.md`
+
+Edit `docs/community/presets.md` to add or update a row in the Community
+Presets table.
+
+### For a new preset
+
+Insert a new row in **alphabetical order by preset name**:
+
+```
+| <Name> | <Description> | <N> templates, <N> commands | <Requires> | [<repo-name>](<repository-url>) |
+```
+
+For the Requires column:
+- Use `—` if no extensions are required
+- List required extension names if any (e.g., `AIDE extension`)
+
+If the preset provides scripts, include them: `<N> templates, <N> commands, <N> scripts`
+
+### For an update
+
+Find the existing row and update any changed fields in-place.
+
+## Step 6 — Create Pull Request
+
+Create a pull request with the changes. Use this branch naming convention:
+
+- **New preset:** `add-<preset-id>-preset`
+- **Update:** `update-<preset-id>-preset`
+
+### Commit message
+
+For a new preset:
+```
+Add <Name> preset to community catalog
+
+Add <id> preset submitted by @<issue-author> to:
+- presets/catalog.community.json (alphabetical order)
+- docs/community/presets.md community presets table
+
+Closes #<issue-number>
+```
+
+For an update:
+```
+Update <Name> preset to v<version>
+
+Update <id> preset submitted by @<issue-author>:
+- presets/catalog.community.json (version, download_url, etc.)
+- docs/community/presets.md community presets table
+
+Closes #<issue-number>
+```
+
+### PR description
+
+Include:
+- A summary of what changed
+- Validation results (all checks passed)
+- `Closes #${{ github.event.issue.number }}`
+- `cc @<issue-author>` — mention the submitter
+
+## Important Rules
+
+- **Alphabetical order matters** — entries must be sorted by ID in the JSON and
+  by name in the docs table
+- **Always validate JSON** after editing — a trailing comma or missing brace
+  will break the catalog
+- **Use `Closes` not `Fixes`** — `Closes #N` is the correct keyword for
+  submission issues
+- **Preserve `created_at` on updates** — keep the original value; only update
+  `updated_at`
+- **Do not modify any other files** — only `presets/catalog.community.json`
+  and `docs/community/presets.md`

.github/workflows/codeql.yml

@@ -22,11 +22,11 @@ jobs:
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
 
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@68bde559dea0fdcac2102bfdf6230c5f70eb485e # v4
+        uses: github/codeql-action/init@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v4
         with:
           languages: ${{ matrix.language }}
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@68bde559dea0fdcac2102bfdf6230c5f70eb485e # v4
+        uses: github/codeql-action/analyze@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v4
         with:
           category: "/language:${{ matrix.language }}"

.github/workflows/lint.yml

@@ -13,6 +13,28 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
+        with:
+          fetch-depth: 1
+
+      - name: Run git diff --check
+        shell: bash
+        env:
+          EVENT_NAME: ${{ github.event_name }}
+          PR_BASE_SHA: ${{ github.event.pull_request.base.sha }}
+          PUSH_BEFORE_SHA: ${{ github.event.before }}
+          GITHUB_SHA: ${{ github.sha }}
+        run: |
+          set -euo pipefail
+
+          if [ "$EVENT_NAME" = "pull_request" ]; then
+            git fetch --no-tags --depth=1 origin "+${PR_BASE_SHA}:refs/checks/pr-base"
+            git diff --check refs/checks/pr-base HEAD
+          elif [ "$PUSH_BEFORE_SHA" = "0000000000000000000000000000000000000000" ]; then
+            git diff-tree --check --no-commit-id --root -r "$GITHUB_SHA"
+          else
+            git fetch --no-tags --depth=1 origin "+${PUSH_BEFORE_SHA}:refs/checks/push-before"
+            git diff --check refs/checks/push-before HEAD
+          fi
 
       - name: Run markdownlint-cli2
         uses: DavidAnson/markdownlint-cli2-action@ded1f9488f68a970bc66ea5619e13e9b52e601cd # v23

.github/workflows/stale.yml

@@ -14,7 +14,7 @@ jobs:
   stale:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/stale@b5d41d4e1d5dceea10e7104786b73624c18a190f # v10
+      - uses: actions/stale@eb5cf3af3ac0a1aa4c9c45633dd1ae542a27a899 # v10
         with:
           # Days of inactivity before an issue or PR becomes stale
           days-before-stale: 150

AGENTS.md

@@ -379,6 +379,33 @@ Implementation: Extends `YamlIntegration` (parallel to `TomlIntegration`):
 4. Uses `yaml.safe_dump()` for header fields to ensure proper escaping
 5. Sets `context_file = "AGENTS.md"` so the base setup manages the Spec Kit context section there
 
+## Branch Naming Convention
+
+Branches follow one of two patterns depending on whether an issue exists:
+
+```
+<type>/<number>-<short-slug>   # when an issue is created first
+<type>/<short-slug>            # when no issue exists (PR-only changes)
+```
+
+When an issue exists, include its number immediately after the prefix — this is what makes branches traceable. For small or self-contained changes that go straight to a PR without a tracking issue, omit the number.
+
+| Prefix | When to use | Example |
+|---|---|---|
+| `feat/` | New features | `feat/2342-workflow-cli-alignment` |
+| `fix/` | Bug fixes | `fix/2653-paths-only-validation` |
+| `docs/` | Documentation changes | `docs/2677-branch-naming-convention`, `docs/update-landing-stats` |
+| `community/` | Community catalog additions | `community/2492-add-mde-extension` |
+| `chore/` | Maintenance, tooling, CI | `chore/2366-editorconfig` |
+
+**Rules:**
+
+1. Include the issue number when one exists — this is what makes branches traceable
+2. Use kebab-case for the slug
+3. Keep the slug short — enough to identify the work without looking up the issue
+
+---
+
 ## Common Pitfalls
 
 1. **Using shorthand keys for CLI-based integrations**: For CLI-based integrations (`requires_cli: True`), the `key` must match the executable name (e.g., `"cursor-agent"` not `"cursor"`). `shutil.which(key)` is used for CLI tool checks — mismatches require special-case mappings. IDE-based integrations (`requires_cli: False`) are not subject to this constraint.

CHANGELOG.md

@@ -2,6 +2,106 @@
 
 <!-- insert new changelog below this comment -->
 
+## [0.8.18] - 2026-05-29
+
+### Changed
+
+- Add support for SPECKIT_WORKFLOW_RUN_ID override (#2742)
+- feat: support SPECKIT_INTEGRATION_<KEY>_EXECUTABLE env var (#2743)
+- chore(deps): bump github/gh-aw-actions from 0.74.8 to 0.77.0 (#2754)
+- chore(deps): bump github/codeql-action from 4.35.5 to 4.36.0 (#2753)
+- fix: disable no-op issue reporting for catalog submission workflows (#2748)
+- Add confirmation prompt for URL-based extension installs (#2745)
+- fix: restrict community submission workflows to labeled event only (#2741)
+- feat(integrations): support SPECIFY_<KEY>_EXTRA_ARGS env var for agent subprocess flags (#2596)
+- chore: release 0.8.17, begin 0.8.18.dev0 development (#2737)
+
+## [0.8.17] - 2026-05-28
+
+### Changed
+
+- docs: consolidate Community sections in README (#2736)
+- Fix shared script command hints for integration separators (#2627)
+- docs: update security-governance preset to v0.4.0 (#2703)
+- feat(agy): enhance Google Antigravity CLI integration (#2689)
+- Fix --dev extension agent symlinks (#2554)
+- Share skills hook note post-processing (#2679)
+- feat: add Hermes Agent integration (with review fixes) (#2651)
+- Update Superpowers Implementation Bridge to v0.7.0 (#2732)
+- chore: release 0.8.16, begin 0.8.17.dev0 development (#2729)
+
+## [0.8.16] - 2026-05-27
+
+### Changed
+
+- docs: update landing page stats and branch naming convention (#2727)
+- feat(workflows): expose {{ context.run_id }} template variable (#2664)
+- fix: resolve __SPECKIT_COMMAND_*__ refs in preset skill rendering (#2717) (#2718)
+- Add Workflow Preset to community catalog (#2725)
+- fix: paths-only skips branch validation, setup-plan preserves existing plan (#2672)
+- docs: fix broken pipx homepage URLs to point to pipx.pypa.io (#2670)
+- Update Architecture Guard extension to v1.8.9 (#2723)
+- Re-validate spec quality checklist after clarify updates spec (#2715)
+- chore: release 0.8.15, begin 0.8.16.dev0 development (#2722)
+
+## [0.8.15] - 2026-05-27
+
+### Changed
+
+- Update Fiction Book Writing preset to v1.8.1 (#2714)
+- chore: update memorylint and superb to 1.4.0 (#2690)
+- fix: promote post-execution hook dispatch to H2 with directive language (#2713)
+- Add Token Budget extension to community catalog (#2712)
+- fix: create skills directory on demand during extension/preset install (#2711)
+- fix: PS 5.1 compat — replace non-ASCII chars in shipped PowerShell scripts (#2709)
+- docs: update security-governance preset to v0.3.0 (#2676)
+- Update README.md (#2675)
+- chore: release 0.8.14, begin 0.8.15.dev0 development (#2706)
+
+## [0.8.14] - 2026-05-26
+
+### Changed
+
+- Add util for windows sub-process (#2598)
+- refactor: create commands/ package and move init handler (PR-4/8) (#2615)
+- Add Product Spec Extension to community catalog (#2705)
+- fix init-options speckit version refresh (#2647)
+- chore(deps): bump github/gh-aw-actions from 0.74.8 to 0.74.9 (#2658)
+- docs: add branch naming convention to AGENTS.md and CONTRIBUTING.md (#2678)
+- chore(deps): bump actions/stale from 10.2.0 to 10.3.0 (#2657)
+- chore(deps): bump github/codeql-action from 4.35.4 to 4.35.5 (#2656)
+- chore: release 0.8.13, begin 0.8.14.dev0 development (#2669)
+
+## [0.8.13] - 2026-05-21
+
+### Changed
+
+- fix: while/do-while loop condition reads stale iteration-0 step output (#2662)
+- docs: fix directory hierarchy in README examples (#2639)
+- fix(catalogs): reject boolean priority in extension and preset catalog readers (#2589)
+- Update Agent Governance extension to v1.2.0 (#2659)
+- Add agentic workflows for community catalog submissions (#2655)
+- feat: add self-check tip to check output (#2574)
+- fix(cli): clarify exception diagnostics (#2602)
+- ci: add diff whitespace check (#2572)
+- chore: release 0.8.12, begin 0.8.13.dev0 development (#2648)
+
+## [0.8.12] - 2026-05-20
+
+### Changed
+
+- fix(codex): inject dot-to-hyphen hook command note in Codex skills (#2503)
+- Update Squad Bridge extension to v1.3.0 (#2645)
+- Update Superpowers Implementation Bridge extension to v0.5.0 (#2644)
+- Add Team Assign extension to community catalog (#2642)
+- refactor: migrate extension catalog stack parsing to shared base (#2576)
+- Update Architecture Workflow extension to v1.1.0 (#2588)
+-  fix(workflow): support integration: auto to follow project's initialized AI (#2421)
+- Add Superpowers Implementation Bridge extension to community catalog (#2586)
+- Add Interactive HTML Preview extension to community catalog (#2585)
+- chore: release 0.8.11, begin 0.8.12.dev0 development (#2584)
+- Update Agent Governance extension to v1.1.0 (#2583)
+
 ## [0.8.11] - 2026-05-15
 
 ### Changed

CONTRIBUTING.md

@@ -38,7 +38,7 @@ On [GitHub Codespaces](https://github.com/features/codespaces) it's even simpler
 1. Fork and clone the repository
 1. Configure and install the dependencies: `uv sync --extra test`
 1. Make sure the CLI works on your machine: `uv run specify --help`
-1. Create a new branch: `git checkout -b my-branch-name`
+1. Create a new branch: `git checkout -b <type>/<number>-<short-slug>` (see [Branch naming](#branch-naming) below)
 1. Make your change, add tests, and make sure everything still works
 1. Test the CLI functionality with a sample project if relevant
 1. Push to your fork and submit a pull request
@@ -55,6 +55,20 @@ Here are a few things you can do that will increase the likelihood of your pull
 - Write a [good commit message](http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html).
 - Test your changes with the Spec-Driven Development workflow to ensure compatibility.
 
+### Branch naming
+
+We recommend naming branches as `<type>/<number>-<short-slug>`, where `<number>` is the issue or PR number (whichever comes first) and `<type>` is one of:
+
+| Prefix | When to use | Example |
+|---|---|---|
+| `feat/` | New features | `feat/2342-workflow-cli-alignment` |
+| `fix/` | Bug fixes | `fix/2653-paths-only-validation` |
+| `docs/` | Documentation changes | `docs/2677-branch-naming-convention` |
+| `community/` | Community catalog additions | `community/2492-add-mde-extension` |
+| `chore/` | Maintenance, tooling, CI | `chore/2366-editorconfig` |
+
+Including the issue or PR number makes branches traceable — especially useful since the project uses squash merges and `git branch --merged` won't detect merged branches. If you start with a PR (no issue), use the PR number once it's assigned.
+
 ## Development workflow
 
 When working on spec-kit:

README.md

@@ -22,10 +22,7 @@
 - [🤔 What is Spec-Driven Development?](#-what-is-spec-driven-development)
 - [⚡ Get Started](#-get-started)
 - [📽️ Video Overview](#️-video-overview)
-- [🧩 Community Extensions](#-community-extensions)
-- [🎨 Community Presets](#-community-presets)
-- [🚶 Community Walkthroughs](#-community-walkthroughs)
-- [🛠️ Community Friends](#️-community-friends)
+- [🌍 Community](#-community)
 - [🤖 Supported AI Coding Agent Integrations](#-supported-ai-coding-agent-integrations)
 - [🔧 Specify CLI Reference](#-specify-cli-reference)
 - [🧩 Making Spec Kit Your Own: Extensions & Presets](#-making-spec-kit-your-own-extensions--presets)
@@ -35,7 +32,7 @@
 - [🔧 Prerequisites](#-prerequisites)
 - [📖 Learn More](#-learn-more)
 - [📋 Detailed Process](#-detailed-process)
-- [ Support](#-support)
+- [💬 Support](#-support)
 - [🙏 Acknowledgements](#-acknowledgements)
 - [📄 License](#-license)
 
@@ -112,31 +109,19 @@ Want to see Spec Kit in action? Watch our [video overview](https://www.youtube.c
 
 [![Spec Kit video header](/media/spec-kit-video-header.jpg)](https://www.youtube.com/watch?v=a9eR1xsfvHg&pp=0gcJCckJAYcqIYzv)
 
-## 🧩 Community Extensions
+## 🌍 Community
 
-Community-contributed extensions add new commands, hooks, and capabilities to Spec Kit. See the full list on the [Community Extensions](https://github.github.io/spec-kit/community/extensions.html) page.
+Explore community-contributed resources on the [Spec Kit docs site](https://github.github.io/spec-kit/):
+
+- [Extensions](https://github.github.io/spec-kit/community/extensions.html) — commands, hooks, and capabilities
+- [Presets](https://github.github.io/spec-kit/community/presets.html) — template and terminology overrides
+- [Walkthroughs](https://github.github.io/spec-kit/community/walkthroughs.html) — end-to-end SDD scenarios
+- [Friends](https://github.github.io/spec-kit/community/friends.html) — projects that extend or build on Spec Kit
 
 > [!NOTE]
-> Community extensions are independently created and maintained by their respective authors. Maintainers only verify that catalog entries are complete and correctly formatted — they do **not review, audit, endorse, or support the extension code itself**. Review extension source code before installation and use at your own discretion.
+> Community contributions are independently created and maintained by their respective authors. Review source code before installation and use at your own discretion.
 
-To submit your own extension, see the [Extension Publishing Guide](extensions/EXTENSION-PUBLISHING-GUIDE.md).
-
-## 🎨 Community Presets
-
-Community-contributed presets customize how Spec Kit behaves — overriding templates, commands, and terminology without changing any tooling. See the full list on the [Community Presets](https://github.github.io/spec-kit/community/presets.html) page.
-
-> [!NOTE]
-> Community presets are third-party contributions and are not maintained by the Spec Kit team. Review them carefully before use, and see the docs page above for the full disclaimer.
-
-To submit your own preset, see the [Presets Publishing Guide](presets/PUBLISHING.md).
-
-## 🚶 Community Walkthroughs
-
-See Spec-Driven Development in action across different scenarios with community-contributed walkthroughs; find the full list on the [Community Walkthroughs](https://github.github.io/spec-kit/community/walkthroughs.html) page.
-
-## 🛠️ Community Friends
-
-Community projects that extend, visualize, or build on Spec Kit. See the full list on the [Community Friends](https://github.github.io/spec-kit/community/friends.html) page.
+Want to contribute? See the [Extension Publishing Guide](extensions/EXTENSION-PUBLISHING-GUIDE.md) or the [Presets Publishing Guide](presets/PUBLISHING.md).
 
 ## 🤖 Supported AI Coding Agent Integrations
 
@@ -206,7 +191,7 @@ specify extension add <extension-name>
 
 For example, extensions could add Jira integration, post-implementation code review, V-Model test traceability, or project health diagnostics.
 
-See the [Extensions reference](https://github.github.io/spec-kit/reference/extensions.html) for the full command guide. Browse the [community extensions](#-community-extensions) above for what's available.
+See the [Extensions reference](https://github.github.io/spec-kit/reference/extensions.html) for the full command guide. Browse the [community extensions](https://github.github.io/spec-kit/community/extensions.html) for what's available.
 
 ### Presets — Customize Existing Workflows
 
@@ -281,7 +266,7 @@ Our research and experimentation focus on:
 
 - **Linux/macOS/Windows**
 - [Supported](#-supported-ai-coding-agent-integrations) AI coding agent.
-- [uv](https://docs.astral.sh/uv/) for package management (recommended) or [pipx](https://pypa.github.io/pipx/) for persistent installation
+- [uv](https://docs.astral.sh/uv/) for package management (recommended) or [pipx](https://pipx.pypa.io/) for persistent installation
 - [Python 3.11+](https://www.python.org/downloads/)
 - [Git](https://git-scm.com/downloads)
 
@@ -400,23 +385,24 @@ The produced specification should contain a set of user stories and functional r
 At this stage, your project folder contents should resemble the following:
 
 ```text
-└── .specify
-    ├── memory
-    │  └── constitution.md
-    ├── scripts
-    │  └── bash
-    │      ├── check-prerequisites.sh
-    │      ├── common.sh
-    │      ├── create-new-feature.sh
-    │      ├── setup-plan.sh
-    │      └── setup-tasks.sh
-    ├── specs
-    │  └── 001-create-taskify
-    │      └── spec.md
-    └── templates
-        ├── plan-template.md
-        ├── spec-template.md
-        └── tasks-template.md
+.
+├── .specify
+│   ├── memory
+│   │   └── constitution.md
+│   ├── scripts
+│   │   └── bash
+│   │       ├── check-prerequisites.sh
+│   │       ├── common.sh
+│   │       ├── create-new-feature.sh
+│   │       ├── setup-plan.sh
+│   │       └── setup-tasks.sh
+│   └── templates
+│       ├── plan-template.md
+│       ├── spec-template.md
+│       └── tasks-template.md
+└── specs
+    └── 001-create-taskify
+        └── spec.md
 ```
 
 ### **STEP 3:** Functional specification clarification (required before planning)
@@ -463,30 +449,31 @@ The output of this step will include a number of implementation detail documents
 ```text
 .
 ├── CLAUDE.md
-├── memory
-│  └── constitution.md
-├── scripts
-│  └── bash
-│      ├── check-prerequisites.sh
-│      ├── common.sh
-│      ├── create-new-feature.sh
-│      ├── setup-plan.sh
-│      └── setup-tasks.sh
-├── specs
-│  └── 001-create-taskify
-│      ├── contracts
-│      │  ├── api-spec.json
-│      │  └── signalr-spec.md
-│      ├── data-model.md
-│      ├── plan.md
-│      ├── quickstart.md
-│      ├── research.md
-│      └── spec.md
-└── templates
-    ├── CLAUDE-template.md
-    ├── plan-template.md
-    ├── spec-template.md
-    └── tasks-template.md
+├── .specify
+│   ├── memory
+│   │   └── constitution.md
+│   ├── scripts
+│   │   └── bash
+│   │       ├── check-prerequisites.sh
+│   │       ├── common.sh
+│   │       ├── create-new-feature.sh
+│   │       ├── setup-plan.sh
+│   │       └── setup-tasks.sh
+│   └── templates
+│       ├── CLAUDE-template.md
+│       ├── plan-template.md
+│       ├── spec-template.md
+│       └── tasks-template.md
+└── specs
+    └── 001-create-taskify
+        ├── contracts
+        │   ├── api-spec.json
+        │   └── signalr-spec.md
+        ├── data-model.md
+        ├── plan.md
+        ├── quickstart.md
+        ├── research.md
+        └── spec.md
 ```
 
 Check the `research.md` document to ensure that the right tech stack is used, based on your instructions. You can ask Claude Code to refine it if any of the components stand out, or even have it check the locally-installed version of the platform/framework you want to use (e.g., .NET).
@@ -579,7 +566,7 @@ Once the implementation is complete, test the application and resolve any runtim
 
 ---
 
-##  Support
+## 💬 Support
 
 For support, please open a [GitHub issue](https://github.com/github/spec-kit/issues/new). We welcome bug reports, feature requests, and questions about using Spec-Driven Development.
 

docs/community/extensions.md

@@ -23,11 +23,11 @@ The following community-contributed extensions are available in [`catalog.commun
 | Extension | Purpose | Category | Effect | URL |
 |-----------|---------|----------|--------|-----|
 | Agent Assign | Assign specialized Claude Code agents to spec-kit tasks for targeted execution | `process` | Read+Write | [spec-kit-agent-assign](https://github.com/xymelon/spec-kit-agent-assign) |
-| Agent Governance | Project-local agent governance memory and context projection | `process` | Read+Write | [spec-kit-agent-governance](https://github.com/bigsmartben/spec-kit-agent-governance) |
+| Agent Governance | Generate agent-platform repository governance files from Spec Kit metadata | `process` | Read+Write | [spec-kit-agent-governance](https://github.com/bigsmartben/spec-kit-agent-governance) |
 | AI-Driven Engineering (AIDE) | A structured 7-step workflow for building new projects from scratch with AI assistants — from vision through implementation | `process` | Read+Write | [aide](https://github.com/mnriem/spec-kit-extensions/tree/main/aide) |
 | API Evolve | Managed API contract evolution — breaking-change detection, semver enforcement, deprecation orchestration, and lifecycle gates across REST, GraphQL, and gRPC | `process` | Read+Write | [spec-kit-api-evolve](https://github.com/Quratulain-bilal/spec-kit-api-evolve) |
 | Architect Impact Previewer | Predicts architectural impact, complexity, and risks of proposed changes before implementation. | `visibility` | Read-only | [spec-kit-architect-preview](https://github.com/UmmeHabiba1312/spec-kit-architect-preview) |
-| Architecture Guard | Continuous architecture governance for AI-assisted development. Reviews specs, plans, and code for architecture drift, producing structured refactor tasks and evolution proposals. | `process` | Read+Write | [spec-kit-architecture-guard](https://github.com/DyanGalih/spec-kit-architecture-guard) |
+| Architecture Guard | Framework-agnostic architecture review extension for validating implementation against governance and architecture constitutions, detecting architectural drift, and generating non-blocking refactor tasks | `process` | Read+Write | [spec-kit-architecture-guard](https://github.com/DyanGalih/spec-kit-architecture-guard) |
 | Architecture Workflow | Generate or reverse project-level 4+1 architecture view artifacts and synthesis | `docs` | Read+Write | [spec-kit-arch](https://github.com/bigsmartben/spec-kit-arch) |
 | Archive Extension | Archive merged features into main project memory. | `docs` | Read+Write | [spec-kit-archive](https://github.com/stn1slv/spec-kit-archive) |
 | Azure DevOps Integration | Sync user stories and tasks to Azure DevOps work items using OAuth authentication | `integration` | Read+Write | [spec-kit-azure-devops](https://github.com/pragya247/spec-kit-azure-devops) |
@@ -78,6 +78,7 @@ The following community-contributed extensions are available in [`catalog.commun
 | PR Bridge | Auto-generate pull request descriptions, checklists, and summaries from spec artifacts | `process` | Read-only | [spec-kit-pr-bridge-](https://github.com/Quratulain-bilal/spec-kit-pr-bridge-) |
 | Presetify | Create and validate presets and preset catalogs | `process` | Read+Write | [presetify](https://github.com/mnriem/spec-kit-extensions/tree/main/presetify) |
 | Product Forge | Full product lifecycle from research to release — portfolio, lite mode, monorepo, optional V-Model | `process` | Read+Write | [speckit-product-forge](https://github.com/VaiYav/speckit-product-forge) |
+| Product Spec Extension | Generates PRFAQ, Lean PRD, stakeholder summaries, and technical designs from engineering specs | `docs` | Read+Write | [spec-kit-product](https://github.com/d0whc3r/spec-kit-product) |
 | Project Health Check | Diagnose a Spec Kit project and report health issues across structure, agents, features, scripts, extensions, and git | `visibility` | Read-only | [spec-kit-doctor](https://github.com/KhawarHabibKhan/spec-kit-doctor) |
 | Project Status | Show current SDD workflow progress — active feature, artifact status, task completion, workflow phase, and extensions summary | `visibility` | Read-only | [spec-kit-status](https://github.com/KhawarHabibKhan/spec-kit-status) |
 | QA Testing Extension | Systematic QA testing with browser-driven or CLI-based validation of acceptance criteria from spec | `code` | Read-only | [spec-kit-qa](https://github.com/arunt14/spec-kit-qa) |
@@ -115,6 +116,7 @@ The following community-contributed extensions are available in [`catalog.commun
 | Team Assign | Assign tasks.md items to human engineers, split into subtasks, and generate a per-engineer workboard | `process` | Read+Write | [spec-kit-team-assign](https://github.com/tarunkumarbhati/spec-kit-team-assign) |
 | Time Machine | Retroactively apply the full SDD workflow to existing codebases — analyse, spec, and ship feature-by-feature | `process` | Read+Write | [spec-kit-time-machine](https://github.com/teeyo/spec-kit-time-machine) |
 | TinySpec | Lightweight single-file workflow for small tasks — skip the heavy multi-step SDD process | `process` | Read+Write | [spec-kit-tinyspec](https://github.com/Quratulain-bilal/spec-kit-tinyspec) |
+| Token Budget | Reduces LLM token consumption in Spec Kit workflows: compact artifacts in-place, scope per-phase reading, suppress prose padding, and report token usage | `process` | Read+Write | [spec-kit-token-budget](https://github.com/tinesoft/spec-kit-token-budget) |
 | Token Consumption Analyzer | Captures, analyzes, and compares token consumption across SDD workflows | `visibility` | Read-only | [spec-kit-token-analyzer](https://github.com/coderandhiker/spec-kit-token-analyzer) |
 | V-Model Extension Pack | Enforces V-Model paired generation of development specs and test specs with full traceability | `docs` | Read+Write | [spec-kit-v-model](https://github.com/leocamello/spec-kit-v-model) |
 | Verify Extension | Post-implementation quality gate that validates implemented code against specification artifacts | `code` | Read-only | [spec-kit-verify](https://github.com/ismaelJimenez/spec-kit-verify) |

docs/community/presets.md

@@ -15,7 +15,7 @@ The following community-contributed presets customize how Spec Kit behaves — o
 | Claude AskUserQuestion | Upgrades `/speckit.clarify` and `/speckit.checklist` on Claude Code from Markdown-table prompts to the native AskUserQuestion picker, with a recommended option and reasoning on every question | 2 commands | — | [spec-kit-preset-claude-ask-questions](https://github.com/0xrafasec/spec-kit-preset-claude-ask-questions) |
 | Cross-Platform Governance | Adds Bash/PowerShell parity, dry-run/WhatIf parity, Unix man-page expectations, PowerShell comment-based help, and Verb-Noun Cmdlet discipline | 8 templates, 3 commands | — | [spec-kit-preset-cross-platform-governance](https://github.com/hindermath/spec-kit-preset-cross-platform-governance) |
 | Explicit Task Dependencies | Adds explicit `(depends on T###)` dependency declarations and an Execution Wave DAG to tasks.md for parallel scheduling | 1 template, 1 command | — | [spec-kit-preset-explicit-task-dependencies](https://github.com/Quratulain-bilal/spec-kit-preset-explicit-task-dependencies) |
-| Fiction Book Writing | It adapts the Spec-Driven Development workflow for storytelling to create books or audiobooks (with annotations) in 12 languages: features become story elements, specs become story briefs, plans become story structures, and tasks become scene-by-scene writing tasks. Supports single and multi-POV, all major plot structure frameworks, and two style modes: an author voice sample or humanized AI prose. Supports interactive elements like brainstorming, interview, roleplay and extras like statistics, cover builder and bio command. Export with templates for KDP, D2D etc. | 22 templates, 27 commands, 2 scripts | — | [speckit-preset-fiction-book-writing](https://github.com/adaumann/speckit-preset-fiction-book-writing) |
+| Fiction Book Writing | It adapts the Spec-Driven Development workflow for storytelling to create books or audiobooks (with annotations) in 12 languages: features become story elements, specs become story briefs, plans become story structures, and tasks become scene-by-scene writing tasks. Supports single and multi-POV, all major plot structure frameworks, and two style modes: an author voice sample or humanized AI prose principles. Supports interactive elements like brainstorming, interview, roleplay and extras like statistics, cover builder and bio command. Export with templates for KDP, D2D etc. | 25 templates, 33 commands, 2 scripts | — | [speckit-preset-fiction-book-writing](https://github.com/adaumann/speckit-preset-fiction-book-writing) |
 | Game Narrative Writing | Spec-Driven Development for interactive game narrative pre-production for video games. Authors write in a portable generic format, Twine/Sugarcube (.twee) or Ink (.ink). Covers choice-IF, visual novels, and branching dialogue. Supports Tier 1 mechanic hooks (flag, counter, inventory, timer, trust, currency, npc_state, ending_condition), multi-ending design, series carry-over variable registry, and NPC-focused character architecture. | 22 templates, 36 commands, 2 scripts | — | [speckit-preset-game-narrative-writing](https://github.com/adaumann/speckit-preset-game-narrative-writing) |
 | iSAQB Architecture Governance | Adds general iSAQB/CPSA-F and arc42 architecture governance: goals, context, building blocks, runtime and deployment views, quality scenarios, ADRs, risks, and technical debt | 13 templates, 3 commands | — | [spec-kit-preset-isaqb-architecture-governance](https://github.com/hindermath/spec-kit-preset-isaqb-architecture-governance) |
 | Jira Issue Tracking | Overrides `speckit.taskstoissues` to create Jira epics, stories, and tasks instead of GitHub Issues via Atlassian MCP tools | 1 command | — | [spec-kit-preset-jira](https://github.com/luno/spec-kit-preset-jira) |
@@ -23,9 +23,10 @@ The following community-contributed presets customize how Spec Kit behaves — o
 | Multi-Repo Branching | Coordinates feature branch creation across multiple git repositories (independent repos and submodules) during plan and tasks phases | 2 commands | — | [spec-kit-preset-multi-repo-branching](https://github.com/sakitA/spec-kit-preset-multi-repo-branching) |
 | Pirate Speak (Full) | Transforms all Spec Kit output into pirate speak — specs become "Voyage Manifests", plans become "Battle Plans", tasks become "Crew Assignments" | 6 templates, 9 commands | — | [spec-kit-presets](https://github.com/mnriem/spec-kit-presets) |
 | Screenwriting | Spec-Driven Development for screenwriting/scriptwriting/tutorials: feature films, television (pilot, episode, limited series), and stage plays. Adapts the Spec Kit workflow to screenplay craft — slug lines, action lines, act breaks, beat sheets, and industry-standard pitch documents. Supports three-act, Save the Cat, TV pilot, network episode, cable/streaming episode, and stage-play structural frameworks. Export to Fountain, FTX, PDF | 26 templates, 32 commands, 1 script | — | [speckit-preset-screenwriting](https://github.com/adaumann/speckit-preset-screenwriting) |
-| Security Governance | Adds secure development governance: memory-safe-language preference, secure code generation, NIST SSDF, CWE Top 25, OWASP ASVS, SBOM/VEX/SLSA, OpenSSF Scorecard, and EU CRA applicability | 12 templates, 3 commands | — | [spec-kit-preset-security-governance](https://github.com/hindermath/spec-kit-preset-security-governance) |
+| Security Governance | Adds secure development governance: memory-safe-language preference, language-specific secure-coding profiles, NIST SSDF, CWE Top 25, OWASP ASVS, SBOM/AI-SBOM, VEX/SLSA, OpenSSF Scorecard, G7/BSI AI-SBOM target evidence, and EU CRA applicability | 12 templates, 3 commands | — | [spec-kit-preset-security-governance](https://github.com/hindermath/spec-kit-preset-security-governance) |
 | Spec2Cloud | Spec-driven workflow tuned for shipping to Azure: spec → plan → tasks → implement → deploy | 5 templates, 8 commands | — | [spec2cloud](https://github.com/Azure-Samples/Spec2Cloud) |
 | Table of Contents Navigation | Adds a navigable Table of Contents to generated spec.md, plan.md, and tasks.md documents | 3 templates, 3 commands | — | [spec-kit-preset-toc-navigation](https://github.com/Quratulain-bilal/spec-kit-preset-toc-navigation) |
 | VS Code Ask Questions | Enhances the clarify command to use `vscode/askQuestions` for batched interactive questioning. | 1 command | — | [spec-kit-presets](https://github.com/fdcastel/spec-kit-presets) |
+| Workflow Preset | Behavior-first specification, design artifacts, and agent-native handoff orchestration — adds requirement-phase behavior drafts, formal BDD/UIF/behavior contracts, optional design artifacts, and scoped implementation handoffs with Core Agent, Vertical Planner Agent, and Worker Agent modes | 23 templates, 7 commands | — | [spec-kit-workflow-preset](https://github.com/bigsmartben/spec-kit-workflow-preset) |
 
 To build and publish your own preset, see the [Presets Publishing Guide](https://github.com/github/spec-kit/blob/main/presets/PUBLISHING.md).

docs/index.md

@@ -43,7 +43,7 @@ Run `specify init` with your agent of choice and Spec Kit sets up the right comm
 
 ### Make it your own
 
-<span class="pillar-stat">91 community extensions</span> (50+ authors), <span class="pillar-stat">18 presets</span>, and growing. Tune the core process with presets, extend it with extensions, orchestrate it with workflows, or replace it entirely. Build and publish your own.
+<span class="pillar-stat">105 community extensions</span> (60+ authors), <span class="pillar-stat">22 presets</span>, and growing. Tune the core process with presets, extend it with extensions, orchestrate it with workflows, or replace it entirely. Build and publish your own.
 
 Including entirely different SDD processes:
 
@@ -82,7 +82,7 @@ Community extensions like CI Guard and Architecture Guard add compliance gates a
 
 <div class="stats-grid">
   <div class="stat-item">
-    <span class="stat-number">96K+</span>
+    <span class="stat-number">106K+</span>
     <span class="stat-label">GitHub stars</span>
   </div>
   <div class="stat-item">
@@ -94,11 +94,11 @@ Community extensions like CI Guard and Architecture Guard add compliance gates a
     <span class="stat-label">Integrations</span>
   </div>
   <div class="stat-item">
-    <span class="stat-number">91</span>
+    <span class="stat-number">105</span>
     <span class="stat-label">Extensions</span>
   </div>
   <div class="stat-item">
-    <span class="stat-number">18</span>
+    <span class="stat-number">22</span>
     <span class="stat-label">Presets</span>
   </div>
   <div class="stat-item">
@@ -150,3 +150,5 @@ specify init my-project --integration copilot
 Ready to start? Follow the [Quick Start Guide](quickstart.md).
 
 </div>
+
+<p class="text-end small text-body-secondary">Last updated: May 27, 2026</p>

docs/install/pipx.md

@@ -1,6 +1,6 @@
 # Installing with pipx
 
-[pipx](https://pypa.github.io/pipx/) is a tool for installing Python CLI applications in isolated environments. It does not require [uv](https://docs.astral.sh/uv/).
+[pipx](https://pipx.pypa.io/) is a tool for installing Python CLI applications in isolated environments. It does not require [uv](https://docs.astral.sh/uv/).
 
 ## Install Specify CLI
 

docs/installation.md

@@ -4,7 +4,7 @@
 
 - **Linux/macOS** (or Windows; PowerShell scripts now supported without WSL)
 - AI coding agent: [Claude Code](https://www.anthropic.com/claude-code), [GitHub Copilot](https://code.visualstudio.com/), [Codebuddy CLI](https://www.codebuddy.ai/cli), [Gemini CLI](https://github.com/google-gemini/gemini-cli), or [Pi Coding Agent](https://pi.dev)
-- [uv](https://docs.astral.sh/uv/) for package management (recommended) or [pipx](https://pypa.github.io/pipx/) for persistent installation
+- [uv](https://docs.astral.sh/uv/) for package management (recommended) or [pipx](https://pipx.pypa.io/) for persistent installation
 - [Python 3.11+](https://www.python.org/downloads/)
 - [Git](https://git-scm.com/downloads)
 

docs/reference/core.md

@@ -69,6 +69,8 @@ specify check
 
 Checks that required tools are available on your system: `git` and any CLI-based AI coding agents. IDE-based agents are skipped since they don't require a CLI tool.
 
+This command stays offline. If a command behaves like an older Spec Kit version or an expected CLI feature is missing, run `specify self check` to check whether your local CLI is behind the latest release.
+
 ## Version Information
 
 ```bash

docs/upgrade.md

@@ -388,6 +388,14 @@ Only Spec Kit infrastructure files:
 
 ### "CLI upgrade doesn't seem to work"
 
+If a command behaves like an older Spec Kit version, first check for local CLI drift:
+
+```bash
+specify self check
+```
+
+`specify check` is an offline environment scan; `specify self check` is the CLI version lookup.
+
 Verify the installation:
 
 ```bash

extensions/README.md

@@ -76,7 +76,7 @@ specify extension add <extension-name> --from https://github.com/org/spec-kit-ex
 
 🔍 **Browse and search community extensions on the [Community Extensions website](https://speckit-community.github.io/extensions/).**
 
-See the [Community Extensions](../README.md#-community-extensions) section in the main README for the full list of available community-contributed extensions.
+See the [Community Extensions](https://github.github.io/spec-kit/community/extensions.html) page for the full list of available community-contributed extensions.
 
 For the raw catalog data, see [`catalog.community.json`](catalog.community.json).
 

extensions/catalog.community.json

@@ -1,6 +1,6 @@
 {
   "schema_version": "1.0",
-  "updated_at": "2026-05-20T00:00:00Z",
+  "updated_at": "2026-05-28T00:00:00Z",
   "catalog_url": "https://raw.githubusercontent.com/github/spec-kit/main/extensions/catalog.community.json",
   "extensions": {
     "aide": {
@@ -71,10 +71,10 @@
     "agent-governance": {
       "name": "Agent Governance",
       "id": "agent-governance",
-      "description": "Project-local agent governance memory and context projection.",
+      "description": "Generate agent-platform repository governance files from Spec Kit metadata.",
       "author": "bigben",
-      "version": "1.1.0",
-      "download_url": "https://github.com/bigsmartben/spec-kit-agent-governance/archive/refs/tags/v1.1.0.zip",
+      "version": "1.2.0",
+      "download_url": "https://github.com/bigsmartben/spec-kit-agent-governance/archive/refs/tags/v1.2.0.zip",
       "repository": "https://github.com/bigsmartben/spec-kit-agent-governance",
       "homepage": "https://github.com/bigsmartben/spec-kit-agent-governance",
       "documentation": "https://github.com/bigsmartben/spec-kit-agent-governance/blob/main/README.md",
@@ -84,8 +84,8 @@
         "speckit_version": ">=0.8.0",
         "tools": [
           {
-            "name": "python3",
-            "required": false
+            "name": "uv",
+            "required": true
           }
         ]
       },
@@ -103,7 +103,7 @@
       "downloads": 0,
       "stars": 0,
       "created_at": "2026-05-14T00:00:00Z",
-      "updated_at": "2026-05-15T00:00:00Z"
+      "updated_at": "2026-05-21T00:00:00Z"
     },
     "agent-orchestrator": {
       "name": "Intelligent Agent Orchestrator",
@@ -240,10 +240,10 @@
     "architecture-guard": {
       "name": "Architecture Guard",
       "id": "architecture-guard",
-      "description": "Continuous architecture governance for AI-assisted development. Reviews specs, plans, and code for architecture drift, producing structured refactor tasks and evolution proposals.",
+      "description": "Framework-agnostic architecture review extension for validating implementation against governance and architecture constitutions, detecting architectural drift, and generating non-blocking refactor tasks.",
       "author": "DyanGalih",
-      "version": "1.8.4",
-      "download_url": "https://github.com/DyanGalih/spec-kit-architecture-guard/archive/refs/tags/v1.8.4.zip",
+      "version": "1.8.9",
+      "download_url": "https://github.com/DyanGalih/spec-kit-architecture-guard/archive/refs/tags/v1.8.9.zip",
       "repository": "https://github.com/DyanGalih/spec-kit-architecture-guard",
       "homepage": "https://github.com/DyanGalih/spec-kit-architecture-guard",
       "documentation": "https://github.com/DyanGalih/spec-kit-architecture-guard/blob/main/README.md",
@@ -258,17 +258,18 @@
       },
       "tags": [
         "architecture",
-        "governance",
-        "drift-detection",
+        "spec-kit",
+        "review",
         "refactor",
-        "monolithic",
-        "microservices"
+        "workflow",
+        "governance",
+        "guardrails"
       ],
       "verified": false,
       "downloads": 0,
       "stars": 0,
       "created_at": "2026-05-05T07:26:00Z",
-      "updated_at": "2026-05-11T14:58:00Z"
+      "updated_at": "2026-05-27T00:00:00Z"
     },
     "archive": {
       "name": "Archive Extension",
@@ -1646,8 +1647,8 @@
       "id": "memorylint",
       "description": "Agent memory governance tool: Automatically audits and fixes boundary conflicts between AGENTS.md and the constitution.",
       "author": "RbBtSn0w",
-      "version": "1.3.0",
-      "download_url": "https://github.com/RbBtSn0w/spec-kit-extensions/releases/download/memorylint-v1.3.0/memorylint.zip",
+      "version": "1.4.0",
+      "download_url": "https://github.com/RbBtSn0w/spec-kit-extensions/releases/download/memorylint-v1.4.0/memorylint.zip",
       "repository": "https://github.com/RbBtSn0w/spec-kit-extensions",
       "homepage": "https://github.com/RbBtSn0w/spec-kit-extensions/tree/main/memorylint",
       "documentation": "https://github.com/RbBtSn0w/spec-kit-extensions/blob/main/memorylint/README.md",
@@ -1657,8 +1658,8 @@
         "speckit_version": ">=0.5.1"
       },
       "provides": {
-        "commands": 1,
-        "hooks": 1
+        "commands": 2,
+        "hooks": 3
       },
       "tags": [
         "memory",
@@ -1671,7 +1672,7 @@
       "downloads": 0,
       "stars": 0,
       "created_at": "2026-04-09T00:00:00Z",
-      "updated_at": "2026-04-16T13:10:26Z"
+      "updated_at": "2026-05-24T01:06:49Z"
     },
     "multi-model-review": {
       "name": "Multi-Model Review",
@@ -1945,6 +1946,38 @@
       "created_at": "2026-05-15T00:00:00Z",
       "updated_at": "2026-05-15T00:00:00Z"
     },
+    "product": {
+      "name": "Product Spec Extension",
+      "id": "product",
+      "description": "Generates PRFAQ, Lean PRD, stakeholder summaries, and technical designs from engineering specs.",
+      "author": "spec-kit-product contributors",
+      "version": "0.1.3",
+      "download_url": "https://github.com/d0whc3r/spec-kit-product/releases/download/v0.1.3/product-0.1.3.zip",
+      "repository": "https://github.com/d0whc3r/spec-kit-product",
+      "homepage": "https://github.com/d0whc3r/spec-kit-product",
+      "documentation": "https://github.com/d0whc3r/spec-kit-product/blob/main/README.md",
+      "changelog": "https://github.com/d0whc3r/spec-kit-product/blob/main/CHANGELOG.md",
+      "license": "MIT",
+      "requires": {
+        "speckit_version": ">=0.2.0"
+      },
+      "provides": {
+        "commands": 4,
+        "hooks": 6
+      },
+      "tags": [
+        "product",
+        "spec",
+        "prd",
+        "design",
+        "documentation"
+      ],
+      "verified": false,
+      "downloads": 0,
+      "stars": 0,
+      "created_at": "2026-05-26T00:00:00Z",
+      "updated_at": "2026-05-26T00:00:00Z"
+    },
     "product-forge": {
       "name": "Product Forge",
       "id": "product-forge",
@@ -2617,8 +2650,8 @@
       "id": "speckit-superpowers-bridge",
       "description": "Thin orchestrator between Spec Kit (design) and Superpowers (implementation). Cross-agent.",
       "author": "lihan3238",
-      "version": "0.5.0",
-      "download_url": "https://github.com/lihan3238/speckit-superpowers-bridge/releases/download/v0.5.0/speckit-superpowers-bridge-v0.5.0.zip",
+      "version": "0.7.0",
+      "download_url": "https://github.com/lihan3238/speckit-superpowers-bridge/releases/download/v0.7.0/speckit-superpowers-bridge-v0.7.0.zip",
       "repository": "https://github.com/lihan3238/speckit-superpowers-bridge",
       "homepage": "https://github.com/lihan3238/speckit-superpowers-bridge",
       "documentation": "https://github.com/lihan3238/speckit-superpowers-bridge#readme",
@@ -2659,7 +2692,7 @@
       "downloads": 0,
       "stars": 0,
       "created_at": "2026-05-15T00:00:00Z",
-      "updated_at": "2026-05-20T00:00:00Z"
+      "updated_at": "2026-05-28T00:00:00Z"
     },
     "speckit-utils": {
       "name": "SDD Utilities",
@@ -2862,8 +2895,8 @@
       "id": "superb",
       "description": "Orchestrates obra/superpowers skills within the spec-kit SDD workflow. Thin bridge commands delegate to superpowers' authoritative SKILL.md files at runtime (with graceful fallback), while bridge-original commands provide spec-kit-native value. Eight commands cover the full lifecycle: intent clarification, TDD enforcement, task review, verification, critique, systematic debugging, branch completion, and review response. Hook-bound commands fire automatically; standalone commands are invoked when needed.",
       "author": "rbbtsn0w",
-      "version": "1.3.0",
-      "download_url": "https://github.com/RbBtSn0w/spec-kit-extensions/releases/download/superpowers-bridge-v1.3.0/superpowers-bridge.zip",
+      "version": "1.4.0",
+      "download_url": "https://github.com/RbBtSn0w/spec-kit-extensions/releases/download/superpowers-bridge-v1.4.0/superpowers-bridge.zip",
       "repository": "https://github.com/RbBtSn0w/spec-kit-extensions",
       "homepage": "https://github.com/RbBtSn0w/spec-kit-extensions",
       "documentation": "https://github.com/RbBtSn0w/spec-kit-extensions/blob/main/superpowers-bridge/README.md",
@@ -2881,7 +2914,7 @@
       },
       "provides": {
         "commands": 8,
-        "hooks": 4
+        "hooks": 3
       },
       "tags": [
         "methodology",
@@ -2898,7 +2931,7 @@
       "downloads": 0,
       "stars": 0,
       "created_at": "2026-03-30T00:00:00Z",
-      "updated_at": "2026-04-16T14:08:23Z"
+      "updated_at": "2026-05-24T01:07:34Z"
     },
     "superpowers-bridge": {
       "name": "Superpowers Bridge",
@@ -3128,6 +3161,48 @@
       "created_at": "2026-05-01T00:00:00Z",
       "updated_at": "2026-05-01T00:00:00Z"
     },
+    "token-budget": {
+      "name": "Token Budget",
+      "id": "token-budget",
+      "description": "Reduces LLM token consumption in Spec Kit workflows: compact artifacts in-place, scope per-phase reading, suppress prose padding, and report token usage.",
+      "author": "Tine Kondo",
+      "version": "1.0.1",
+      "download_url": "https://github.com/tinesoft/spec-kit-token-budget/archive/refs/tags/v1.0.1.zip",
+      "repository": "https://github.com/tinesoft/spec-kit-token-budget",
+      "homepage": "https://github.com/tinesoft/spec-kit-token-budget",
+      "documentation": "https://github.com/tinesoft/spec-kit-token-budget/blob/main/README.md",
+      "changelog": "https://github.com/tinesoft/spec-kit-token-budget/blob/main/CHANGELOG.md",
+      "license": "MIT",
+      "requires": {
+        "speckit_version": ">=0.1.0",
+        "tools": [
+          {
+            "name": "python3",
+            "required": false
+          },
+          {
+            "name": "rtk",
+            "required": false
+          }
+        ]
+      },
+      "provides": {
+        "commands": 4,
+        "hooks": 6
+      },
+      "tags": [
+        "tokens",
+        "budget",
+        "context",
+        "efficiency",
+        "cost-optimization"
+      ],
+      "verified": false,
+      "downloads": 0,
+      "stars": 0,
+      "created_at": "2026-05-26T00:00:00Z",
+      "updated_at": "2026-05-26T00:00:00Z"
+    },
     "v-model": {
       "name": "V-Model Extension Pack",
       "id": "v-model",

extensions/git/commands/speckit.git.initialize.md

@@ -35,7 +35,7 @@ Replace the script to add project-specific Git initialization steps:
 ## Output
 
 On success:
-- `✓ Git repository initialized`
+- `[OK] Git repository initialized`
 
 ## Graceful Degradation
 

extensions/git/scripts/powershell/auto-commit.ps1

@@ -115,7 +115,7 @@ if (Test-Path $configFile) {
         }
     }
 } else {
-    # No config file — auto-commit disabled by default
+    # No config file -- auto-commit disabled by default
     exit 0
 }
 

extensions/git/scripts/powershell/git-common.ps1

@@ -1,6 +1,6 @@
 #!/usr/bin/env pwsh
 # Git-specific common functions for the git extension.
-# Extracted from scripts/powershell/common.ps1 — contains only git-specific
+# Extracted from scripts/powershell/common.ps1 -- contains only git-specific
 # branch validation and detection logic.
 
 function Test-HasGit {

extensions/git/scripts/powershell/initialize-repo.ps1

@@ -1,7 +1,7 @@
 #!/usr/bin/env pwsh
 # Git extension: initialize-repo.ps1
 # Initialize a Git repository with an initial commit.
-# Customizable — replace this script to add .gitignore templates,
+# Customizable -- replace this script to add .gitignore templates,
 # default branch config, git-flow, LFS, signing, etc.
 $ErrorActionPreference = 'Stop'
 
@@ -66,4 +66,4 @@ try {
     exit 1
 }
 
-Write-Host "✓ Git repository initialized"
+Write-Host "[OK] Git repository initialized"

integrations/catalog.json

@@ -272,6 +272,15 @@
       "author": "spec-kit-core",
       "repository": "https://github.com/github/spec-kit",
       "tags": ["cli"]
+    },
+    "hermes": {
+      "id": "hermes",
+      "name": "Hermes Agent",
+      "version": "1.0.0",
+      "description": "Hermes Agent skills-based integration by Nous Research",
+      "author": "spec-kit-core",
+      "repository": "https://github.com/github/spec-kit",
+      "tags": ["cli", "skills"]
     }
   }
 }

presets/catalog.community.json

@@ -1,6 +1,6 @@
 {
   "schema_version": "1.0",
-  "updated_at": "2026-05-05T10:00:00Z",
+  "updated_at": "2026-05-27T00:00:00Z",
   "catalog_url": "https://raw.githubusercontent.com/github/spec-kit/main/presets/catalog.community.json",
   "presets": {
     "a11y-governance": {
@@ -222,11 +222,11 @@
     "fiction-book-writing": {
       "name": "Fiction Book Writing",
       "id": "fiction-book-writing",
-      "version": "1.7.0",
-      "description": "Spec-Driven Development for novel and long-form fiction. 27 AI commands from idea to submission: story bible governance, 9 POV modes, all major plot structure frameworks, scene-by-scene drafting with quality gates, audiobook pipeline (SSML/ElevenLabs), cover design, sensitivity review, pacing and prose statistics, and pandoc-based export to DOCX/EPUB/LaTeX. Two style modes: author voice sample extraction or humanized-AI prose with 5 craft profiles. 12 languages supported. Support for offline semantic search.",
+      "version": "1.8.1",
+      "description": "Spec-Driven Development for novel and long-form fiction. 33 AI commands from idea to submission: story bible governance, 9 POV modes, all major plot structure frameworks, scene-by-scene drafting with quality gates, audiobook pipeline (SSML/ElevenLabs), cover design, sensitivity review, pacing and prose statistics, and pandoc-based export to DOCX/EPUB/LaTeX. Two style modes: author voice sample extraction or humanized-AI prose with 5 craft profiles. 12 languages supported. Support for offline semantic search.",
       "author": "Andreas Daumann",
       "repository": "https://github.com/adaumann/speckit-preset-fiction-book-writing",
-      "download_url": "https://github.com/adaumann/speckit-preset-fiction-book-writing/archive/refs/tags/v1.7.0.zip",
+      "download_url": "https://github.com/adaumann/speckit-preset-fiction-book-writing/archive/refs/tags/v1.8.1.zip",
       "homepage": "https://github.com/adaumann/speckit-preset-fiction-book-writing",
       "documentation": "https://github.com/adaumann/speckit-preset-fiction-book-writing/blob/main/fiction-book-writing/README.md",
       "license": "MIT",
@@ -234,8 +234,8 @@
         "speckit_version": ">=0.5.0"
       },
       "provides": {
-        "templates": 22,
-        "commands": 27,
+        "templates": 25,
+        "commands": 33,
         "scripts": 2
       },
       "tags": [
@@ -254,7 +254,7 @@
         "language-support"
       ],
       "created_at": "2026-04-09T08:00:00Z",
-      "updated_at": "2026-04-27T08:00:00Z"
+      "updated_at": "2026-05-24T08:00:00Z"
     },
     "game-narrative-writing": {
       "name": "Game Narrative Writing",
@@ -472,11 +472,11 @@
     "security-governance": {
       "name": "Security Governance",
       "id": "security-governance",
-      "version": "0.2.0",
-      "description": "Adds secure development governance, MSL preference, ASVS verification, supply-chain transparency, and EU CRA awareness.",
+      "version": "0.4.0",
+      "description": "Adds memory-safe-language preference, language-specific secure coding profiles, ASVS verification, SBOM/AI-SBOM supply-chain transparency, and EU Cyber Resilience Act awareness.",
       "author": "Thorsten Hindermann",
       "repository": "https://github.com/hindermath/spec-kit-preset-security-governance",
-      "download_url": "https://github.com/hindermath/spec-kit-preset-security-governance/archive/refs/tags/v0.2.0.zip",
+      "download_url": "https://github.com/hindermath/spec-kit-preset-security-governance/archive/refs/tags/v0.4.0.zip",
       "homepage": "https://github.com/hindermath/spec-kit-preset-security-governance",
       "documentation": "https://github.com/hindermath/spec-kit-preset-security-governance/blob/main/README.md",
       "license": "MIT",
@@ -491,11 +491,28 @@
         "security",
         "governance",
         "msl",
+        "ssdf",
         "asvs",
-        "supply-chain"
+        "supply-chain",
+        "sbom",
+        "ai-sbom",
+        "vex",
+        "slsa",
+        "cwe-top-25",
+        "secure-coding",
+        "rust",
+        "go",
+        "swift",
+        "java",
+        "kotlin",
+        "python",
+        "typescript",
+        "g7",
+        "bsi",
+        "cra"
       ],
       "created_at": "2026-04-27T00:00:00Z",
-      "updated_at": "2026-04-27T00:00:00Z"
+      "updated_at": "2026-05-26T00:00:00Z"
     },
     "spec2cloud": {
       "name": "Spec2Cloud",
@@ -572,6 +589,34 @@
         "clarify",
         "interactive"
       ]
+    },
+    "workflow-preset": {
+      "name": "Workflow Preset",
+      "id": "workflow-preset",
+      "version": "1.2.0",
+      "description": "Behavior-first specification, design artifacts, and agent-native handoff orchestration.",
+      "author": "bigsmartben",
+      "repository": "https://github.com/bigsmartben/spec-kit-workflow-preset",
+      "download_url": "https://github.com/bigsmartben/spec-kit-workflow-preset/archive/refs/tags/v1.2.0.zip",
+      "homepage": "https://github.com/bigsmartben/spec-kit-workflow-preset",
+      "documentation": "https://github.com/bigsmartben/spec-kit-workflow-preset/blob/main/README.md",
+      "license": "MIT",
+      "requires": {
+        "speckit_version": ">=0.8.10.dev0"
+      },
+      "provides": {
+        "templates": 23,
+        "commands": 7
+      },
+      "tags": [
+        "behavior",
+        "bdd",
+        "planning",
+        "implementation",
+        "handoff"
+      ],
+      "created_at": "2026-05-27T00:00:00Z",
+      "updated_at": "2026-05-27T00:00:00Z"
     }
   }
 }

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "specify-cli"
-version = "0.8.12.dev0"
+version = "0.8.18"
 description = "Specify CLI, part of GitHub Spec Kit. A tool to bootstrap your projects for Spec-Driven Development (SDD)."
 requires-python = ">=3.11"
 dependencies = [

scripts/bash/check-prerequisites.sh

@@ -78,13 +78,12 @@ done
 SCRIPT_DIR="$(CDPATH="" cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 source "$SCRIPT_DIR/common.sh"
 
-# Get feature paths and validate branch
+# Get feature paths
 _paths_output=$(get_feature_paths) || { echo "ERROR: Failed to resolve feature paths" >&2; exit 1; }
 eval "$_paths_output"
 unset _paths_output
-check_feature_branch "$CURRENT_BRANCH" "$HAS_GIT" || exit 1
 
-# If paths-only mode, output paths and exit (support JSON + paths-only combined)
+# If paths-only mode, output paths and exit (no validation)
 if $PATHS_ONLY; then
     if $JSON_MODE; then
         # Minimal JSON paths payload (no validation performed)
@@ -112,23 +111,26 @@ if $PATHS_ONLY; then
     exit 0
 fi
 
+# Validate branch name
+check_feature_branch "$CURRENT_BRANCH" "$HAS_GIT" || exit 1
+
 # Validate required directories and files
 if [[ ! -d "$FEATURE_DIR" ]]; then
     echo "ERROR: Feature directory not found: $FEATURE_DIR" >&2
-    echo "Run /speckit.specify first to create the feature structure." >&2
+    echo "Run __SPECKIT_COMMAND_SPECIFY__ first to create the feature structure." >&2
     exit 1
 fi
 
 if [[ ! -f "$IMPL_PLAN" ]]; then
     echo "ERROR: plan.md not found in $FEATURE_DIR" >&2
-    echo "Run /speckit.plan first to create the implementation plan." >&2
+    echo "Run __SPECKIT_COMMAND_PLAN__ first to create the implementation plan." >&2
     exit 1
 fi
 
 # Check for tasks.md if required
 if $REQUIRE_TASKS && [[ ! -f "$TASKS" ]]; then
     echo "ERROR: tasks.md not found in $FEATURE_DIR" >&2
-    echo "Run /speckit.tasks first to create the task list." >&2
+    echo "Run __SPECKIT_COMMAND_TASKS__ first to create the task list." >&2
     exit 1
 fi
 

scripts/bash/common.sh

@@ -186,7 +186,7 @@ read_feature_json_feature_directory() {
 }
 
 # Returns 0 when .specify/feature.json lists feature_directory that exists as a directory
-# and matches the resolved active FEATURE_DIR (so /speckit.plan can skip git branch pattern checks).
+# and matches the resolved active FEATURE_DIR (so __SPECKIT_COMMAND_PLAN__ can skip git branch pattern checks).
 # Delegates parsing to read_feature_json_feature_directory, which is safe under `set -e`.
 feature_json_matches_feature_dir() {
     local repo_root="$1"
@@ -262,7 +262,7 @@ get_feature_paths() {
 
     # Resolve feature directory.  Priority:
     #   1. SPECIFY_FEATURE_DIRECTORY env var (explicit override)
-    #   2. .specify/feature.json "feature_directory" key (persisted by /speckit.specify)
+    #   2. .specify/feature.json "feature_directory" key (persisted by __SPECKIT_COMMAND_SPECIFY__)
     #   3. Branch-name-based prefix lookup (legacy fallback)
     local feature_dir
     if [[ -n "${SPECIFY_FEATURE_DIRECTORY:-}" ]]; then
@@ -642,4 +642,3 @@ except Exception:
     printf '%s' "$content"
     return 0
 }
-

scripts/bash/setup-plan.sh

@@ -40,15 +40,31 @@ fi
 # Ensure the feature directory exists
 mkdir -p "$FEATURE_DIR"
 
-# Copy plan template if it exists
-TEMPLATE=$(resolve_template "plan-template" "$REPO_ROOT") || true
-if [[ -n "$TEMPLATE" ]] && [[ -f "$TEMPLATE" ]]; then
-    cp "$TEMPLATE" "$IMPL_PLAN"
-    echo "Copied plan template to $IMPL_PLAN"
+# Copy plan template if plan doesn't already exist
+if [[ -f "$IMPL_PLAN" ]]; then
+    if $JSON_MODE; then
+        echo "Plan already exists at $IMPL_PLAN, skipping template copy" >&2
+    else
+        echo "Plan already exists at $IMPL_PLAN, skipping template copy"
+    fi
 else
-    echo "Warning: Plan template not found"
-    # Create a basic plan file if template doesn't exist
-    touch "$IMPL_PLAN"
+    TEMPLATE=$(resolve_template "plan-template" "$REPO_ROOT") || true
+    if [[ -n "$TEMPLATE" ]] && [[ -f "$TEMPLATE" ]]; then
+        cp "$TEMPLATE" "$IMPL_PLAN"
+        if $JSON_MODE; then
+            echo "Copied plan template to $IMPL_PLAN" >&2
+        else
+            echo "Copied plan template to $IMPL_PLAN"
+        fi
+    else
+        if $JSON_MODE; then
+            echo "Warning: Plan template not found" >&2
+        else
+            echo "Warning: Plan template not found"
+        fi
+        # Create a basic plan file if template doesn't exist
+        touch "$IMPL_PLAN"
+    fi
 fi
 
 # Output results

scripts/bash/setup-tasks.sh

@@ -35,13 +35,13 @@ fi
 
 if [[ ! -f "$IMPL_PLAN" ]]; then
     echo "ERROR: plan.md not found in $FEATURE_DIR" >&2
-    echo "Run /speckit.plan first to create the implementation plan." >&2
+    echo "Run __SPECKIT_COMMAND_PLAN__ first to create the implementation plan." >&2
     exit 1
 fi
 
 if [[ ! -f "$FEATURE_SPEC" ]]; then
     echo "ERROR: spec.md not found in $FEATURE_DIR" >&2
-    echo "Run /speckit.specify first to create the feature structure." >&2
+    echo "Run __SPECKIT_COMMAND_SPECIFY__ first to create the feature structure." >&2
     exit 1
 fi
 

scripts/powershell/check-prerequisites.ps1

@@ -56,14 +56,10 @@ EXAMPLES:
 # Source common functions
 . "$PSScriptRoot/common.ps1"
 
-# Get feature paths and validate branch
+# Get feature paths
 $paths = Get-FeaturePathsEnv
 
-if (-not (Test-FeatureBranch -Branch $paths.CURRENT_BRANCH -HasGit:$paths.HAS_GIT)) { 
-    exit 1 
-}
-
-# If paths-only mode, output paths and exit (support combined -Json -PathsOnly)
+# If paths-only mode, output paths and exit (no validation)
 if ($PathsOnly) {
     if ($Json) {
         [PSCustomObject]@{
@@ -85,23 +81,28 @@ if ($PathsOnly) {
     exit 0
 }
 
+# Validate branch name
+if (-not (Test-FeatureBranch -Branch $paths.CURRENT_BRANCH -HasGit:$paths.HAS_GIT)) {
+    exit 1
+}
+
 # Validate required directories and files
 if (-not (Test-Path $paths.FEATURE_DIR -PathType Container)) {
     Write-Output "ERROR: Feature directory not found: $($paths.FEATURE_DIR)"
-    Write-Output "Run /speckit.specify first to create the feature structure."
+    Write-Output "Run __SPECKIT_COMMAND_SPECIFY__ first to create the feature structure."
     exit 1
 }
 
 if (-not (Test-Path $paths.IMPL_PLAN -PathType Leaf)) {
     Write-Output "ERROR: plan.md not found in $($paths.FEATURE_DIR)"
-    Write-Output "Run /speckit.plan first to create the implementation plan."
+    Write-Output "Run __SPECKIT_COMMAND_PLAN__ first to create the implementation plan."
     exit 1
 }
 
 # Check for tasks.md if required
 if ($RequireTasks -and -not (Test-Path $paths.TASKS -PathType Leaf)) {
     Write-Output "ERROR: tasks.md not found in $($paths.FEATURE_DIR)"
-    Write-Output "Run /speckit.tasks first to create the task list."
+    Write-Output "Run __SPECKIT_COMMAND_TASKS__ first to create the task list."
     exit 1
 }
 

scripts/powershell/common.ps1

@@ -165,7 +165,7 @@ function Test-FeatureBranch {
 }
 
 # True when .specify/feature.json pins an existing feature directory that matches the
-# active FEATURE_DIR from Get-FeaturePathsEnv (so /speckit.plan can skip git branch pattern checks).
+# active FEATURE_DIR from Get-FeaturePathsEnv (so __SPECKIT_COMMAND_PLAN__ can skip git branch pattern checks).
 function Test-FeatureJsonMatchesFeatureDir {
     param(
         [Parameter(Mandatory = $true)][string]$RepoRoot,
@@ -288,7 +288,7 @@ function Get-FeaturePathsEnv {
 
     # Resolve feature directory.  Priority:
     #   1. SPECIFY_FEATURE_DIRECTORY env var (explicit override)
-    #   2. .specify/feature.json "feature_directory" key (persisted by /speckit.specify)
+    #   2. .specify/feature.json "feature_directory" key (persisted by __SPECKIT_COMMAND_SPECIFY__)
     #   3. Branch-name-based prefix lookup (same as scripts/bash/common.sh)
     $featureJson = Join-Path $repoRoot '.specify/feature.json'
     if ($env:SPECIFY_FEATURE_DIRECTORY) {
@@ -336,10 +336,10 @@ function Get-FeaturePathsEnv {
 function Test-FileExists {
     param([string]$Path, [string]$Description)
     if (Test-Path -Path $Path -PathType Leaf) {
-        Write-Output "  ✓ $Description"
+        Write-Output "  [OK] $Description"
         return $true
     } else {
-        Write-Output "  ✗ $Description"
+        Write-Output "  [FAIL] $Description"
         return $false
     }
 }
@@ -347,10 +347,10 @@ function Test-FileExists {
 function Test-DirHasFiles {
     param([string]$Path, [string]$Description)
     if ((Test-Path -Path $Path -PathType Container) -and (Get-ChildItem -Path $Path -ErrorAction SilentlyContinue | Where-Object { -not $_.PSIsContainer } | Select-Object -First 1)) {
-        Write-Output "  ✓ $Description"
+        Write-Output "  [OK] $Description"
         return $true
     } else {
-        Write-Output "  ✗ $Description"
+        Write-Output "  [FAIL] $Description"
         return $false
     }
 }
@@ -591,7 +591,7 @@ except Exception:
 
     if ($layerPaths.Count -eq 0) { return $null }
 
-    # If the top (highest-priority) layer is replace, it wins entirely —
+    # If the top (highest-priority) layer is replace, it wins entirely --
     # lower layers are irrelevant regardless of their strategies.
     if ($layerStrategies[0] -eq 'replace') {
         return (Get-Content $layerPaths[0] -Raw)
@@ -640,4 +640,4 @@ except Exception:
     }
 
     return $content
-}
+}

scripts/powershell/create-new-feature.ps1

@@ -312,7 +312,7 @@ if (-not $DryRun) {
                 if ($AllowExistingBranch) {
                     # If we're already on the branch, continue without another checkout.
                     if ($currentBranch -eq $branchName) {
-                        # Already on the target branch — nothing to do
+                        # Already on the target branch -- nothing to do
                     } else {
                         # Otherwise switch to the existing branch instead of failing.
                         $switchBranchError = git checkout -q $branchName 2>&1 | Out-String

scripts/powershell/setup-plan.ps1

@@ -33,17 +33,25 @@ if (-not (Test-FeatureJsonMatchesFeatureDir -RepoRoot $paths.REPO_ROOT -ActiveFe
 # Ensure the feature directory exists
 New-Item -ItemType Directory -Path $paths.FEATURE_DIR -Force | Out-Null
 
-# Copy plan template if it exists, otherwise note it or create empty file
-$template = Resolve-Template -TemplateName 'plan-template' -RepoRoot $paths.REPO_ROOT
-if ($template -and (Test-Path $template)) { 
-    # Read the template content and write it to the implementation plan file with UTF-8 encoding without BOM
-    $content = [System.IO.File]::ReadAllText($template)
-    $utf8NoBom = New-Object System.Text.UTF8Encoding($false)
-    [System.IO.File]::WriteAllText($paths.IMPL_PLAN, $content, $utf8NoBom)
+# Copy plan template if plan doesn't already exist
+if (Test-Path $paths.IMPL_PLAN -PathType Leaf) {
+    if ($Json) {
+        [Console]::Error.WriteLine("Plan already exists at $($paths.IMPL_PLAN), skipping template copy")
+    } else {
+        Write-Output "Plan already exists at $($paths.IMPL_PLAN), skipping template copy"
+    }
 } else {
-    Write-Warning "Plan template not found"
-    # Create a basic plan file if template doesn't exist
-    New-Item -ItemType File -Path $paths.IMPL_PLAN -Force | Out-Null
+    $template = Resolve-Template -TemplateName 'plan-template' -RepoRoot $paths.REPO_ROOT
+    if ($template -and (Test-Path $template)) {
+        # Read the template content and write it to the implementation plan file with UTF-8 encoding without BOM
+        $content = [System.IO.File]::ReadAllText($template)
+        $utf8NoBom = New-Object System.Text.UTF8Encoding($false)
+        [System.IO.File]::WriteAllText($paths.IMPL_PLAN, $content, $utf8NoBom)
+    } else {
+        Write-Warning "Plan template not found"
+        # Create a basic plan file if template doesn't exist
+        New-Item -ItemType File -Path $paths.IMPL_PLAN -Force | Out-Null
+    }
 }
 
 # Output results

scripts/powershell/setup-tasks.ps1

@@ -28,13 +28,13 @@ if (-not (Test-FeatureJsonMatchesFeatureDir -RepoRoot $paths.REPO_ROOT -ActiveFe
 
 if (-not (Test-Path $paths.IMPL_PLAN -PathType Leaf)) {
     [Console]::Error.WriteLine("ERROR: plan.md not found in $($paths.FEATURE_DIR)")
-    [Console]::Error.WriteLine("Run /speckit.plan first to create the implementation plan.")
+    [Console]::Error.WriteLine("Run __SPECKIT_COMMAND_PLAN__ first to create the implementation plan.")
     exit 1
 }
 
 if (-not (Test-Path $paths.FEATURE_SPEC -PathType Leaf)) {
     [Console]::Error.WriteLine("ERROR: spec.md not found in $($paths.FEATURE_DIR)")
-    [Console]::Error.WriteLine("Run /speckit.specify first to create the feature structure.")
+    [Console]::Error.WriteLine("Run __SPECKIT_COMMAND_SPECIFY__ first to create the feature structure.")
     exit 1
 }
 

src/specify_cli/_agent_config.py

@@ -0,0 +1,45 @@
+"""Agent configuration constants derived from the integration registry."""
+from __future__ import annotations
+
+from typing import Any
+
+
+def _build_agent_config() -> dict[str, dict[str, Any]]:
+    from .integrations import INTEGRATION_REGISTRY
+    config: dict[str, dict[str, Any]] = {}
+    for key, integration in INTEGRATION_REGISTRY.items():
+        if integration.config:
+            config[key] = dict(integration.config)
+    return config
+
+
+AGENT_CONFIG: dict[str, dict[str, Any]] = _build_agent_config()
+
+DEFAULT_INIT_INTEGRATION = "copilot"
+
+AI_ASSISTANT_ALIASES: dict[str, str] = {
+    "kiro": "kiro-cli",
+}
+
+
+def _build_ai_assistant_help() -> str:
+    non_generic_agents = sorted(agent for agent in AGENT_CONFIG if agent != "generic")
+    base_help = (
+        f"AI assistant to use: {', '.join(non_generic_agents)}, "
+        "or generic (requires --ai-commands-dir)."
+    )
+    if not AI_ASSISTANT_ALIASES:
+        return base_help
+    alias_phrases = []
+    for alias, target in sorted(AI_ASSISTANT_ALIASES.items()):
+        alias_phrases.append(f"'{alias}' as an alias for '{target}'")
+    if len(alias_phrases) == 1:
+        aliases_text = alias_phrases[0]
+    else:
+        aliases_text = ", ".join(alias_phrases[:-1]) + " and " + alias_phrases[-1]
+    return base_help + " Use " + aliases_text + "."
+
+
+AI_ASSISTANT_HELP: str = _build_ai_assistant_help()
+
+SCRIPT_TYPE_CHOICES: dict[str, str] = {"sh": "POSIX Shell (bash/zsh)", "ps": "PowerShell"}

src/specify_cli/agents.py

@@ -439,6 +439,7 @@ class CommandRegistrar:
         project_root: Path,
         context_note: str = None,
         _resolved_dir: Path = None,
+        link_outputs: bool = False,
     ) -> List[str]:
         """Register commands for a specific agent.
 
@@ -453,6 +454,9 @@ class CommandRegistrar:
                 only — avoids a second ``_resolve_agent_dir`` call and
                 duplicate deprecation warnings when invoked from
                 ``register_commands_for_all_agents``).
+            link_outputs: If True, write rendered output to a source-local
+                dev cache and symlink the agent command file to it. Falls back
+                to a normal file write when symlinks are unavailable.
 
         Returns:
             List of registered command names
@@ -559,7 +563,15 @@ class CommandRegistrar:
             dest_file = commands_dir / f"{output_name}{agent_config['extension']}"
             self._ensure_inside(dest_file, commands_dir)
             dest_file.parent.mkdir(parents=True, exist_ok=True)
-            dest_file.write_text(output, encoding="utf-8")
+            self._write_registered_output(
+                dest_file,
+                output,
+                source_dir,
+                agent_name,
+                output_name,
+                agent_config["extension"],
+                link_outputs,
+            )
 
             if agent_name == "copilot":
                 self.write_copilot_prompt(project_root, cmd_name)
@@ -625,13 +637,56 @@ class CommandRegistrar:
                 )
                 self._ensure_inside(alias_file, commands_dir)
                 alias_file.parent.mkdir(parents=True, exist_ok=True)
-                alias_file.write_text(alias_output, encoding="utf-8")
+                self._write_registered_output(
+                    alias_file,
+                    alias_output,
+                    source_dir,
+                    agent_name,
+                    alias_output_name,
+                    agent_config["extension"],
+                    link_outputs,
+                )
                 if agent_name == "copilot":
                     self.write_copilot_prompt(project_root, alias)
                 registered.append(alias)
 
         return registered
 
+    @staticmethod
+    def _write_registered_output(
+        dest_file: Path,
+        content: str,
+        source_dir: Path,
+        agent_name: str,
+        output_name: str,
+        extension: str,
+        link_outputs: bool,
+    ) -> None:
+        """Write a rendered agent artifact, optionally as a dev-mode symlink."""
+        if not link_outputs:
+            dest_file.write_text(content, encoding="utf-8")
+            return
+
+        rel_output = Path(f"{output_name}{extension}")
+        cache_root = source_dir / ".specify-dev" / "agent-commands" / agent_name
+        cache_file = cache_root / rel_output
+        CommandRegistrar._ensure_inside(cache_file, cache_root)
+
+        try:
+            cache_file.parent.mkdir(parents=True, exist_ok=True)
+            cache_file.write_text(content, encoding="utf-8")
+            if dest_file.exists() or dest_file.is_symlink():
+                dest_file.unlink()
+            target = os.path.relpath(cache_file, dest_file.parent)
+            os.symlink(target, dest_file)
+        except (OSError, ValueError):
+            # Windows often requires Developer Mode or admin privileges for
+            # symlinks, and relpath can fail across drives. Keep dev installs
+            # functional by falling back to a copy.
+            if dest_file.is_symlink():
+                dest_file.unlink()
+            dest_file.write_text(content, encoding="utf-8")
+
     @staticmethod
     def write_copilot_prompt(project_root: Path, cmd_name: str) -> None:
         """Generate a companion .prompt.md file for a Copilot agent command.
@@ -654,15 +709,28 @@ class CommandRegistrar:
     ) -> Path:
         """Return the agent command directory, falling back to legacy_dir.
 
-        When the canonical directory (``agent_config["dir"]``) does not
-        exist but a ``legacy_dir`` is configured and present on disk,
-        returns the legacy path and emits a deprecation warning advising
-        the user to upgrade.
+        Supports project-relative paths (e.g. ``.claude/skills/``),
+        home-relative paths (e.g. ``~/.hermes/skills``), and absolute
+        paths — the ``agent_config["dir"]`` value is resolved verbatim
+        when absolute or starting with ``~/``, or joined with
+        ``project_root`` when relative.
+
+        When the canonical directory does not exist but a ``legacy_dir``
+        is configured and present on disk, returns the legacy path and
+        emits a deprecation warning advising the user to upgrade.
 
         Integrations that do not declare ``legacy_dir`` get the canonical
         path unconditionally — no fallback, no warning.
         """
-        agent_dir = project_root / agent_config["dir"]
+        dir_str = agent_config["dir"]
+        if dir_str.startswith("~"):
+            # Use Path.home() + remainder instead of expanduser() so tests
+            # that monkeypatch Path.home() can properly isolate the home dir.
+            # expanduser() uses OS env/user lookup and ignores monkeypatches.
+            agent_dir = Path.home() / dir_str[1:].lstrip("/")
+        else:
+            p = Path(dir_str)
+            agent_dir = p if p.is_absolute() else project_root / p
         if not agent_dir.exists():
             legacy = agent_config.get("legacy_dir")
             if legacy:
@@ -687,6 +755,7 @@ class CommandRegistrar:
         source_dir: Path,
         project_root: Path,
         context_note: str = None,
+        link_outputs: bool = False,
     ) -> Dict[str, List[str]]:
         """Register commands for all detected agents in the project.
 
@@ -696,6 +765,8 @@ class CommandRegistrar:
             source_dir: Directory containing command source files
             project_root: Path to project root
             context_note: Custom context comment for markdown output
+            link_outputs: If True, create dev-mode symlinks for rendered
+                command files when supported by the OS.
 
         Returns:
             Dictionary mapping agent names to list of registered commands
@@ -704,6 +775,15 @@ class CommandRegistrar:
 
         self._ensure_configs()
         for agent_name, agent_config in self.AGENT_CONFIGS.items():
+            # Check detect_dir first (project-local marker) if configured,
+            # falling back to the resolved dir for output.  This prevents
+            # global dirs (e.g. ~/.hermes/skills) from causing false
+            # detection in every project.
+            detect_dir_str = agent_config.get("detect_dir")
+            if detect_dir_str:
+                detect_path = project_root / detect_dir_str
+                if not detect_path.exists():
+                    continue
             agent_dir = self._resolve_agent_dir(
                 agent_name, agent_config, project_root,
             )
@@ -718,6 +798,7 @@ class CommandRegistrar:
                         project_root,
                         context_note=context_note,
                         _resolved_dir=agent_dir,
+                        link_outputs=link_outputs,
                     )
                     if registered:
                         results[agent_name] = registered
@@ -733,6 +814,7 @@ class CommandRegistrar:
         source_dir: Path,
         project_root: Path,
         context_note: Optional[str] = None,
+        link_outputs: bool = False,
     ) -> Dict[str, List[str]]:
         """Register commands for all non-skill agents in the project.
 
@@ -746,6 +828,8 @@ class CommandRegistrar:
             source_dir: Directory containing command source files
             project_root: Path to project root
             context_note: Custom context comment for markdown output
+            link_outputs: If True, create dev-mode symlinks for rendered
+                command files when supported by the OS.
 
         Returns:
             Dictionary mapping agent names to list of registered commands
@@ -755,6 +839,11 @@ class CommandRegistrar:
         for agent_name, agent_config in self.AGENT_CONFIGS.items():
             if agent_config.get("extension") == "/SKILL.md":
                 continue
+            detect_dir_str = agent_config.get("detect_dir")
+            if detect_dir_str:
+                detect_path = project_root / detect_dir_str
+                if not detect_path.exists():
+                    continue
             agent_dir = self._resolve_agent_dir(
                 agent_name, agent_config, project_root,
             )
@@ -768,6 +857,7 @@ class CommandRegistrar:
                         project_root,
                         context_note=context_note,
                         _resolved_dir=agent_dir,
+                        link_outputs=link_outputs,
                     )
                     if registered:
                         results[agent_name] = registered
@@ -816,7 +906,7 @@ class CommandRegistrar:
                     cmd_file = (
                         target_dir / f"{output_name}{agent_config['extension']}"
                     )
-                    if cmd_file.exists():
+                    if cmd_file.exists() or cmd_file.is_symlink():
                         cmd_file.unlink()
                         # For SKILL.md agents each command lives in its own
                         # subdirectory (e.g. .agents/skills/speckit-ext-cmd/

src/specify_cli/commands/__init__.py

@@ -0,0 +1,7 @@
+"""CLI command groups extracted from the main application.
+
+Implemented command modules expose a ``register(app)`` function. Placeholder
+modules are import-only anchors for command groups that still live in the main
+application module.
+"""
+from __future__ import annotations

src/specify_cli/commands/extension.py

@@ -0,0 +1,2 @@
+"""specify extension * commands — placeholder for future extraction."""
+from __future__ import annotations

src/specify_cli/commands/init.py

@@ -0,0 +1,743 @@
+"""specify init command."""
+from __future__ import annotations
+
+import os
+import shlex
+import shutil
+import sys
+from pathlib import Path
+from typing import Any
+
+import typer
+from rich.live import Live
+from rich.panel import Panel
+
+from .._agent_config import (
+    AGENT_CONFIG,
+    AI_ASSISTANT_ALIASES,
+    AI_ASSISTANT_HELP,
+    DEFAULT_INIT_INTEGRATION,
+    SCRIPT_TYPE_CHOICES,
+)
+from .._assets import (
+    _locate_bundled_extension,
+    _locate_bundled_preset,
+    _locate_bundled_workflow,
+    get_speckit_version,
+)
+from .._console import StepTracker, console, select_with_arrows, show_banner
+from .._utils import check_tool, init_git_repo, is_git_repo
+
+def _build_integration_equivalent(
+    integration_key: str,
+    ai_commands_dir: str | None = None,
+) -> str:
+    parts = [f"--integration {integration_key}"]
+    if integration_key == "generic" and ai_commands_dir:
+        parts.append(
+            f'--integration-options="--commands-dir {shlex.quote(ai_commands_dir)}"'
+        )
+    return " ".join(parts)
+
+
+def _build_ai_deprecation_warning(
+    integration_key: str,
+    ai_commands_dir: str | None = None,
+) -> str:
+    replacement = _build_integration_equivalent(
+        integration_key,
+        ai_commands_dir=ai_commands_dir,
+    )
+    return (
+        "[bold]--ai[/bold] is deprecated and will no longer be available in version 0.10.0 or later.\n\n"
+        f"Use [bold]{replacement}[/bold] instead."
+    )
+
+
+def _stdin_is_interactive() -> bool:
+    return sys.stdin.isatty()
+
+
+def ensure_constitution_from_template(
+    project_path: Path, tracker: StepTracker | None = None
+) -> None:
+    """Copy constitution template to memory if it doesn't exist."""
+    memory_constitution = project_path / ".specify" / "memory" / "constitution.md"
+    template_constitution = project_path / ".specify" / "templates" / "constitution-template.md"
+
+    if memory_constitution.exists():
+        if tracker:
+            tracker.add("constitution", "Constitution setup")
+            tracker.skip("constitution", "existing file preserved")
+        return
+
+    if not template_constitution.exists():
+        if tracker:
+            tracker.add("constitution", "Constitution setup")
+            tracker.error("constitution", "template not found")
+        return
+
+    try:
+        memory_constitution.parent.mkdir(parents=True, exist_ok=True)
+        shutil.copy2(template_constitution, memory_constitution)
+        if tracker:
+            tracker.add("constitution", "Constitution setup")
+            tracker.complete("constitution", "copied from template")
+        else:
+            console.print("[cyan]Initialized constitution from template[/cyan]")
+    except Exception as e:
+        if tracker:
+            tracker.add("constitution", "Constitution setup")
+            tracker.error("constitution", str(e))
+        else:
+            console.print(f"[yellow]Warning: Could not initialize constitution: {e}[/yellow]")
+
+
+def register(app: typer.Typer) -> None:
+    @app.command()
+    def init(
+        project_name: str = typer.Argument(None, help="Name for your new project directory (optional if using --here, or use '.' for current directory)"),
+        ai_assistant: str = typer.Option(None, "--ai", help=AI_ASSISTANT_HELP),
+        ai_commands_dir: str = typer.Option(None, "--ai-commands-dir", help="Directory for agent command files (required with --ai generic, e.g. .myagent/commands/)"),
+        script_type: str = typer.Option(None, "--script", help="Script type to use: sh or ps"),
+        ignore_agent_tools: bool = typer.Option(False, "--ignore-agent-tools", help="Skip checks for coding agent tools like Claude Code"),
+        no_git: bool = typer.Option(False, "--no-git", help="Skip git repository initialization"),
+        here: bool = typer.Option(False, "--here", help="Initialize project in the current directory instead of creating a new one"),
+        force: bool = typer.Option(False, "--force", help="Force merge/overwrite when using --here (skip confirmation)"),
+        skip_tls: bool = typer.Option(False, "--skip-tls", help="Deprecated (no-op). Previously: skip SSL/TLS verification.", hidden=True),
+        debug: bool = typer.Option(False, "--debug", help="Deprecated. Previously: show verbose diagnostic output; currently only prints additional diagnostic details on failure.", hidden=True),
+        github_token: str = typer.Option(None, "--github-token", help="Deprecated (no-op). Previously: GitHub token for API requests.", hidden=True),
+        ai_skills: bool = typer.Option(False, "--ai-skills", help="Install Prompt.MD templates as agent skills (requires --ai)"),
+        offline: bool = typer.Option(False, "--offline", help="Deprecated (no-op). All scaffolding now uses bundled assets.", hidden=True),
+        preset: str = typer.Option(None, "--preset", help="Install a preset during initialization (by preset ID)"),
+        branch_numbering: str = typer.Option(None, "--branch-numbering", help="Branch numbering strategy: 'sequential' (001, 002, …, 1000, … — expands past 999 automatically) or 'timestamp' (YYYYMMDD-HHMMSS)"),
+        integration: str = typer.Option(None, "--integration", help="Use the new integration system (e.g. --integration copilot). Mutually exclusive with --ai."),
+        integration_options: str = typer.Option(None, "--integration-options", help='Options for the integration (e.g. --integration-options="--commands-dir .myagent/cmds")'),
+    ):
+        """
+        Initialize a new Specify project.
+
+        Project files are scaffolded from assets bundled inside the specify-cli
+        package, so initialization does not need network access and templates
+        match the installed CLI version.
+
+        This command will:
+        1. Check that required tools are installed (git is optional)
+        2. Let you choose your coding agent integration, or default to Copilot
+           in non-interactive sessions
+        3. Install bundled Spec Kit templates, scripts, workflow, and shared
+           project infrastructure
+        4. Initialize a fresh git repository (if not --no-git and no existing repo)
+        5. Set up coding agent integration commands and optional presets
+
+        Examples:
+            specify init my-project
+            specify init my-project --integration claude
+            specify init my-project --integration copilot --no-git
+            specify init --ignore-agent-tools my-project
+            specify init . --integration claude         # Initialize in current directory
+            specify init .                     # Initialize in current directory (interactive integration selection)
+            specify init --here --integration claude    # Alternative syntax for current directory
+            specify init --here --integration codex --integration-options="--skills"
+            specify init --here --integration codebuddy
+            specify init --here --integration vibe      # Initialize with Mistral Vibe support
+            specify init --here
+            specify init --here --force  # Skip confirmation when current directory not empty
+            specify init my-project --integration claude   # Claude installs skills by default
+            specify init --here --integration gemini
+            specify init my-project --integration generic --integration-options="--commands-dir .myagent/commands/"  # Bring your own agent; requires --commands-dir
+            specify init my-project --integration claude --preset healthcare-compliance  # With preset
+        """
+        # Lazy imports to avoid circular dependency — __init__.py imports this module
+        from .. import (
+            _install_shared_infra_or_exit,
+            _parse_integration_options,
+            _print_cli_warning,
+            _write_integration_json,
+            ensure_executable_scripts,
+            save_init_options,
+        )
+        from ..integration_runtime import with_integration_setting as _with_integration_setting
+
+        show_banner()
+        ai_deprecation_warning: str | None = None
+
+        if ai_assistant and ai_assistant.startswith("--"):
+            console.print(f"[red]Error:[/red] Invalid value for --ai: '{ai_assistant}'")
+            console.print("[yellow]Hint:[/yellow] Did you forget to provide a value for --ai?")
+            console.print("[yellow]Example:[/yellow] specify init --integration claude --here")
+            console.print(f"[yellow]Available agents:[/yellow] {', '.join(AGENT_CONFIG.keys())}")
+            raise typer.Exit(1)
+
+        if ai_commands_dir and ai_commands_dir.startswith("--"):
+            console.print(f"[red]Error:[/red] Invalid value for --ai-commands-dir: '{ai_commands_dir}'")
+            console.print("[yellow]Hint:[/yellow] Did you forget to provide a value for --ai-commands-dir?")
+            console.print("[yellow]Example:[/yellow] specify init --integration generic --integration-options=\"--commands-dir .myagent/commands/\"")
+            raise typer.Exit(1)
+
+        if ai_assistant:
+            ai_assistant = AI_ASSISTANT_ALIASES.get(ai_assistant, ai_assistant)
+
+        if integration and ai_assistant:
+            console.print("[red]Error:[/red] --integration and --ai are mutually exclusive")
+            raise typer.Exit(1)
+
+        from ..integrations import INTEGRATION_REGISTRY, get_integration
+        if integration:
+            resolved_integration = get_integration(integration)
+            if not resolved_integration:
+                console.print(f"[red]Error:[/red] Unknown integration: '{integration}'")
+                available = ", ".join(sorted(INTEGRATION_REGISTRY))
+                console.print(f"[yellow]Available integrations:[/yellow] {available}")
+                raise typer.Exit(1)
+            ai_assistant = integration
+        elif ai_assistant:
+            resolved_integration = get_integration(ai_assistant)
+            if not resolved_integration:
+                console.print(f"[red]Error:[/red] Unknown agent '{ai_assistant}'. Choose from: {', '.join(sorted(INTEGRATION_REGISTRY))}")
+                raise typer.Exit(1)
+            ai_deprecation_warning = _build_ai_deprecation_warning(
+                resolved_integration.key,
+                ai_commands_dir=ai_commands_dir,
+            )
+
+        if ai_assistant or integration:
+            if ai_skills:
+                from ..integrations.base import SkillsIntegration as _SkillsCheck
+                if isinstance(resolved_integration, _SkillsCheck):
+                    console.print(
+                        "[dim]Note: --ai-skills is not needed; "
+                        "skills are the default for this integration.[/dim]"
+                    )
+                else:
+                    console.print(
+                        "[dim]Note: --ai-skills has no effect with "
+                        f"{resolved_integration.key}; this integration uses commands, not skills.[/dim]"
+                    )
+            if ai_commands_dir and resolved_integration.key != "generic":
+                console.print(
+                    "[dim]Note: --ai-commands-dir is deprecated; "
+                    'use [bold]--integration generic --integration-options="--commands-dir <dir>"[/bold] instead.[/dim]'
+                )
+
+        if no_git:
+            console.print(
+                "[yellow]⚠️  --no-git is deprecated and will be removed in v0.10.0.[/yellow]\n"
+                "[yellow]The git extension will no longer be enabled by default "
+                "— use the [bold]specify extension[/bold] commands to install or enable the git extension if needed.[/yellow]"
+            )
+
+        if project_name == ".":
+            here = True
+            project_name = None
+
+        if here and project_name:
+            console.print("[red]Error:[/red] Cannot specify both project name and --here flag")
+            raise typer.Exit(1)
+
+        if not here and not project_name:
+            console.print("[red]Error:[/red] Must specify either a project name, use '.' for current directory, or use --here flag")
+            raise typer.Exit(1)
+
+        if ai_skills and not ai_assistant:
+            console.print("[red]Error:[/red] --ai-skills requires --ai to be specified")
+            console.print("[yellow]Usage:[/yellow] specify init <project> --ai <agent> --ai-skills")
+            raise typer.Exit(1)
+
+        BRANCH_NUMBERING_CHOICES = {"sequential", "timestamp"}
+        if branch_numbering and branch_numbering not in BRANCH_NUMBERING_CHOICES:
+            console.print(f"[red]Error:[/red] Invalid --branch-numbering value '{branch_numbering}'. Choose from: {', '.join(sorted(BRANCH_NUMBERING_CHOICES))}")
+            raise typer.Exit(1)
+
+        dir_existed_before = False
+        if here:
+            project_name = Path.cwd().name
+            project_path = Path.cwd()
+            dir_existed_before = True
+
+            existing_items = list(project_path.iterdir())
+            if existing_items:
+                console.print(f"[yellow]Warning:[/yellow] Current directory is not empty ({len(existing_items)} items)")
+                console.print("[yellow]Template files will be merged with existing content and may overwrite existing files[/yellow]")
+                if force:
+                    console.print("[cyan]--force supplied: skipping confirmation and proceeding with merge[/cyan]")
+                else:
+                    response = typer.confirm("Do you want to continue?")
+                    if not response:
+                        console.print("[yellow]Operation cancelled[/yellow]")
+                        raise typer.Exit(0)
+        else:
+            project_path = Path(project_name).resolve()
+            dir_existed_before = project_path.exists()
+            if project_path.exists():
+                if not project_path.is_dir():
+                    console.print(f"[red]Error:[/red] '{project_name}' exists but is not a directory.")
+                    raise typer.Exit(1)
+                existing_items = list(project_path.iterdir())
+                if force:
+                    if existing_items:
+                        console.print(f"[yellow]Warning:[/yellow] Directory '{project_name}' is not empty ({len(existing_items)} items)")
+                        console.print("[yellow]Template files will be merged with existing content and may overwrite existing files[/yellow]")
+                    console.print(f"[cyan]--force supplied: merging into existing directory '[cyan]{project_name}[/cyan]'[/cyan]")
+                else:
+                    error_panel = Panel(
+                        f"Directory already exists: '[cyan]{project_name}[/cyan]'\n"
+                        "Please choose a different project name or remove the existing directory.\n"
+                        "Use [bold]--force[/bold] to merge into the existing directory.",
+                        title="[red]Directory Conflict[/red]",
+                        border_style="red",
+                        padding=(1, 2)
+                    )
+                    console.print()
+                    console.print(error_panel)
+                    raise typer.Exit(1)
+
+        if ai_assistant:
+            if ai_assistant not in AGENT_CONFIG:
+                console.print(f"[red]Error:[/red] Invalid AI assistant '{ai_assistant}'. Choose from: {', '.join(AGENT_CONFIG.keys())}")
+                raise typer.Exit(1)
+            selected_ai = ai_assistant
+        elif not _stdin_is_interactive():
+            console.print(
+                f"[dim]Non-interactive session detected: defaulting to '{DEFAULT_INIT_INTEGRATION}'. "
+                "Use --integration to choose a different agent.[/dim]"
+            )
+            selected_ai = DEFAULT_INIT_INTEGRATION
+        else:
+            ai_choices = {key: config["name"] for key, config in AGENT_CONFIG.items()}
+            selected_ai = select_with_arrows(
+                ai_choices,
+                "Choose your coding agent integration:",
+                DEFAULT_INIT_INTEGRATION,
+            )
+
+        if not ai_assistant:
+            resolved_integration = get_integration(selected_ai)
+            if not resolved_integration:
+                console.print(f"[red]Error:[/red] Unknown agent '{selected_ai}'")
+                raise typer.Exit(1)
+
+        if selected_ai == "generic" and not integration_options:
+            if not ai_commands_dir:
+                console.print("[red]Error:[/red] --ai-commands-dir is required when using --ai generic or --integration generic")
+                console.print('[dim]Example: specify init my-project --integration generic --integration-options="--commands-dir .myagent/commands/"[/dim]')
+                raise typer.Exit(1)
+
+        current_dir = Path.cwd()
+
+        setup_lines = [
+            "[cyan]Specify Project Setup[/cyan]",
+            "",
+            f"{'Project':<15} [green]{project_path.name}[/green]",
+            f"{'Working Path':<15} [dim]{current_dir}[/dim]",
+        ]
+
+        if not here:
+            setup_lines.append(f"{'Target Path':<15} [dim]{project_path}[/dim]")
+
+        console.print(Panel("\n".join(setup_lines), border_style="cyan", padding=(1, 2)))
+
+        should_init_git = False
+        if not no_git:
+            should_init_git = check_tool("git")
+            if not should_init_git:
+                console.print("[yellow]Git not found - will skip repository initialization[/yellow]")
+
+        if not ignore_agent_tools:
+            agent_config = AGENT_CONFIG.get(selected_ai)
+            if agent_config and agent_config["requires_cli"]:
+                install_url = agent_config["install_url"]
+                if not check_tool(selected_ai):
+                    error_panel = Panel(
+                        f"[cyan]{selected_ai}[/cyan] not found\n"
+                        f"Install from: [cyan]{install_url}[/cyan]\n"
+                        f"{agent_config['name']} is required to continue with this project type.\n\n"
+                        "Tip: Use [cyan]--ignore-agent-tools[/cyan] to skip this check",
+                        title="[red]Agent Detection Error[/red]",
+                        border_style="red",
+                        padding=(1, 2)
+                    )
+                    console.print()
+                    console.print(error_panel)
+                    raise typer.Exit(1)
+
+        if script_type:
+            if script_type not in SCRIPT_TYPE_CHOICES:
+                console.print(f"[red]Error:[/red] Invalid script type '{script_type}'. Choose from: {', '.join(SCRIPT_TYPE_CHOICES.keys())}")
+                raise typer.Exit(1)
+            selected_script = script_type
+        else:
+            default_script = "ps" if os.name == "nt" else "sh"
+
+            if _stdin_is_interactive():
+                selected_script = select_with_arrows(SCRIPT_TYPE_CHOICES, "Choose script type (or press Enter)", default_script)
+            else:
+                selected_script = default_script
+
+        console.print(f"[cyan]Selected coding agent integration:[/cyan] {selected_ai}")
+        console.print(f"[cyan]Selected script type:[/cyan] {selected_script}")
+
+        tracker = StepTracker("Initialize Specify Project")
+
+        tracker.add("precheck", "Check required tools")
+        tracker.complete("precheck", "ok")
+        tracker.add("ai-select", "Select coding agent integration")
+        tracker.complete("ai-select", f"{selected_ai}")
+        tracker.add("script-select", "Select script type")
+        tracker.complete("script-select", selected_script)
+
+        tracker.add("integration", "Install integration")
+        tracker.add("shared-infra", "Install shared infrastructure")
+
+        for key, label in [
+            ("chmod", "Ensure scripts executable"),
+            ("constitution", "Constitution setup"),
+            ("git", "Install git extension"),
+            ("workflow", "Install bundled workflow"),
+            ("final", "Finalize"),
+        ]:
+            tracker.add(key, label)
+
+        git_default_notice = False
+
+        with Live(tracker.render(), console=console, refresh_per_second=8, transient=True) as live:
+            tracker.attach_refresh(lambda: live.update(tracker.render()))
+            try:
+                from ..integrations.manifest import IntegrationManifest
+                tracker.start("integration")
+                manifest = IntegrationManifest(
+                    resolved_integration.key, project_path, version=get_speckit_version()
+                )
+
+                integration_parsed_options: dict[str, Any] = {}
+                if ai_commands_dir:
+                    integration_parsed_options["commands_dir"] = ai_commands_dir
+                if ai_skills:
+                    integration_parsed_options["skills"] = True
+                if integration_options:
+                    extra = _parse_integration_options(resolved_integration, integration_options)
+                    if extra:
+                        integration_parsed_options.update(extra)
+
+                resolved_integration.setup(
+                    project_path, manifest,
+                    parsed_options=integration_parsed_options or None,
+                    script_type=selected_script,
+                    raw_options=integration_options,
+                )
+                manifest.save()
+
+                integration_settings = _with_integration_setting(
+                    {},
+                    resolved_integration.key,
+                    resolved_integration,
+                    script_type=selected_script,
+                    raw_options=integration_options,
+                    parsed_options=integration_parsed_options or None,
+                )
+                _write_integration_json(
+                    project_path,
+                    resolved_integration.key,
+                    [resolved_integration.key],
+                    integration_settings,
+                )
+
+                tracker.complete("integration", resolved_integration.config.get("name", resolved_integration.key))
+
+                tracker.start("shared-infra")
+                _install_shared_infra_or_exit(
+                    project_path,
+                    selected_script,
+                    tracker=tracker,
+                    force=force,
+                    invoke_separator=resolved_integration.effective_invoke_separator(integration_parsed_options),
+                )
+                tracker.complete("shared-infra", f"scripts ({selected_script}) + templates")
+
+                ensure_constitution_from_template(project_path, tracker=tracker)
+
+                if not no_git:
+                    tracker.start("git")
+                    git_messages = []
+                    git_has_error = False
+                    if is_git_repo(project_path):
+                        git_messages.append("existing repo detected")
+                    elif should_init_git:
+                        success, error_msg = init_git_repo(project_path, quiet=True)
+                        if success:
+                            git_messages.append("initialized")
+                        else:
+                            git_has_error = True
+                            if error_msg:
+                                sanitized = error_msg.replace('\n', ' ').strip()
+                                git_messages.append(f"init failed: {sanitized[:120]}")
+                            else:
+                                git_messages.append("init failed")
+                    else:
+                        git_messages.append("git not available")
+                    try:
+                        from ..extensions import ExtensionManager
+                        bundled_path = _locate_bundled_extension("git")
+                        if bundled_path:
+                            manager = ExtensionManager(project_path)
+                            if manager.registry.is_installed("git"):
+                                git_messages.append("extension already installed")
+                            else:
+                                manager.install_from_directory(
+                                    bundled_path, get_speckit_version()
+                                )
+                                git_default_notice = True
+                                git_messages.append("extension installed")
+                        else:
+                            git_has_error = True
+                            git_messages.append("bundled extension not found")
+                    except Exception as ext_err:
+                        git_has_error = True
+                        sanitized_ext = str(ext_err).replace('\n', ' ').strip()
+                        git_messages.append(
+                            f"extension install failed: {sanitized_ext[:120]}"
+                        )
+                    summary = "; ".join(git_messages)
+                    if git_has_error:
+                        tracker.error("git", summary)
+                    else:
+                        tracker.complete("git", summary)
+                else:
+                    tracker.skip("git", "--no-git flag")
+
+                try:
+                    bundled_wf = _locate_bundled_workflow("speckit")
+                    if bundled_wf:
+                        from ..workflows.catalog import WorkflowRegistry
+                        from ..workflows.engine import WorkflowDefinition
+                        wf_registry = WorkflowRegistry(project_path)
+                        if wf_registry.is_installed("speckit"):
+                            tracker.complete("workflow", "already installed")
+                        else:
+                            import shutil as _shutil
+                            dest_wf = project_path / ".specify" / "workflows" / "speckit"
+                            dest_wf.mkdir(parents=True, exist_ok=True)
+                            _shutil.copy2(
+                                bundled_wf / "workflow.yml",
+                                dest_wf / "workflow.yml",
+                            )
+                            definition = WorkflowDefinition.from_yaml(dest_wf / "workflow.yml")
+                            wf_registry.add("speckit", {
+                                "name": definition.name,
+                                "version": definition.version,
+                                "description": definition.description,
+                                "source": "bundled",
+                            })
+                            tracker.complete("workflow", "speckit installed")
+                    else:
+                        tracker.skip("workflow", "bundled workflow not found")
+                except Exception as wf_err:
+                    sanitized_wf = str(wf_err).replace('\n', ' ').strip()
+                    tracker.error("workflow", f"install failed: {sanitized_wf[:120]}")
+
+                ensure_executable_scripts(project_path, tracker=tracker)
+
+                init_opts = {
+                    "ai": selected_ai,
+                    "integration": resolved_integration.key,
+                    "branch_numbering": branch_numbering or "sequential",
+                    "context_file": resolved_integration.context_file,
+                    "here": here,
+                    "script": selected_script,
+                    "speckit_version": get_speckit_version(),
+                }
+                from ..integrations.base import SkillsIntegration as _SkillsPersist
+                if isinstance(resolved_integration, _SkillsPersist) or getattr(resolved_integration, "_skills_mode", False):
+                    init_opts["ai_skills"] = True
+                save_init_options(project_path, init_opts)
+
+                if preset:
+                    try:
+                        from ..presets import PresetManager, PresetCatalog, PresetError
+                        preset_manager = PresetManager(project_path)
+                        speckit_ver = get_speckit_version()
+
+                        local_path = Path(preset).resolve()
+                        if local_path.is_dir() and (local_path / "preset.yml").exists():
+                            preset_manager.install_from_directory(local_path, speckit_ver)
+                        else:
+                            bundled_path = _locate_bundled_preset(preset)
+                            if bundled_path:
+                                preset_manager.install_from_directory(bundled_path, speckit_ver)
+                            else:
+                                preset_catalog = PresetCatalog(project_path)
+                                pack_info = preset_catalog.get_pack_info(preset)
+                                if not pack_info:
+                                    console.print(f"[yellow]Warning:[/yellow] Preset '{preset}' not found in catalog. Skipping.")
+                                elif pack_info.get("bundled") and not pack_info.get("download_url"):
+                                    from ..extensions import REINSTALL_COMMAND
+                                    console.print(
+                                        f"[yellow]Warning:[/yellow] Preset '{preset}' is bundled with spec-kit "
+                                        f"but could not be found in the installed package."
+                                    )
+                                    console.print(
+                                        "This usually means the spec-kit installation is incomplete or corrupted."
+                                    )
+                                    console.print(f"Try reinstalling: {REINSTALL_COMMAND}")
+                                else:
+                                    zip_path = None
+                                    try:
+                                        zip_path = preset_catalog.download_pack(preset)
+                                        preset_manager.install_from_zip(zip_path, speckit_ver)
+                                    except PresetError as preset_err:
+                                        _print_cli_warning(
+                                            "install",
+                                            "preset",
+                                            preset,
+                                            preset_err,
+                                            continuing="Continuing without the optional preset.",
+                                        )
+                                    finally:
+                                        if zip_path is not None:
+                                            try:
+                                                zip_path.unlink(missing_ok=True)
+                                            except OSError:
+                                                pass
+                    except Exception as preset_err:
+                        _print_cli_warning(
+                            "install",
+                            "preset",
+                            preset,
+                            preset_err,
+                            continuing="Continuing without the optional preset.",
+                        )
+
+                tracker.complete("final", "project ready")
+            except (typer.Exit, SystemExit):
+                raise
+            except Exception as e:
+                tracker.error("final", str(e))
+                console.print(Panel(f"Initialization failed: {e}", title="Failure", border_style="red"))
+                if debug:
+                    _env_pairs = [
+                        ("Python", sys.version.split()[0]),
+                        ("Platform", sys.platform),
+                        ("CWD", str(Path.cwd())),
+                    ]
+                    _label_width = max(len(k) for k, _ in _env_pairs)
+                    env_lines = [f"{k.ljust(_label_width)} → [bright_black]{v}[/bright_black]" for k, v in _env_pairs]
+                    console.print(Panel("\n".join(env_lines), title="Debug Environment", border_style="magenta"))
+                if not here and project_path.exists() and not dir_existed_before:
+                    shutil.rmtree(project_path)
+                raise typer.Exit(1)
+            finally:
+                pass
+
+        console.print(tracker.render())
+        console.print("\n[bold green]Project ready.[/bold green]")
+
+        agent_config = AGENT_CONFIG.get(selected_ai)
+        if agent_config:
+            agent_folder = ai_commands_dir if selected_ai == "generic" else agent_config["folder"]
+            if agent_folder:
+                security_notice = Panel(
+                    f"Some agents may store credentials, auth tokens, or other identifying and private artifacts in the agent folder within your project.\n"
+                    f"Consider adding [cyan]{agent_folder}[/cyan] (or parts of it) to [cyan].gitignore[/cyan] to prevent accidental credential leakage.",
+                    title="[yellow]Agent Folder Security[/yellow]",
+                    border_style="yellow",
+                    padding=(1, 2)
+                )
+                console.print()
+                console.print(security_notice)
+
+        if ai_deprecation_warning:
+            deprecation_notice = Panel(
+                ai_deprecation_warning,
+                title="[bold red]Deprecation Warning[/bold red]",
+                border_style="red",
+                padding=(1, 2),
+            )
+            console.print()
+            console.print(deprecation_notice)
+
+        if git_default_notice:
+            default_change_notice = Panel(
+                "The git extension is currently enabled by default during [bold]specify init[/bold].\n"
+                "Starting in [bold]v0.10.0[/bold], this will require explicit opt-in.\n"
+                "Use [bold]specify extension add git[/bold] after init when needed.",
+                title="[yellow]Notice: Git Default Changing[/yellow]",
+                border_style="yellow",
+                padding=(1, 2),
+            )
+            console.print()
+            console.print(default_change_notice)
+
+        steps_lines = []
+        if not here:
+            steps_lines.append(f"1. Go to the project folder: [cyan]cd {project_name}[/cyan]")
+            step_num = 2
+        else:
+            steps_lines.append("1. You're already in the project directory!")
+            step_num = 2
+
+        from ..integrations.base import SkillsIntegration as _SkillsInt
+        _is_skills_integration = isinstance(resolved_integration, _SkillsInt) or getattr(resolved_integration, "_skills_mode", False)
+
+        codex_skill_mode = selected_ai == "codex" and (ai_skills or _is_skills_integration)
+        claude_skill_mode = selected_ai == "claude" and (ai_skills or _is_skills_integration)
+        kimi_skill_mode = selected_ai == "kimi"
+        agy_skill_mode = selected_ai == "agy" and _is_skills_integration
+        trae_skill_mode = selected_ai == "trae"
+        cursor_agent_skill_mode = selected_ai == "cursor-agent" and (ai_skills or _is_skills_integration)
+        copilot_skill_mode = selected_ai == "copilot" and _is_skills_integration
+        devin_skill_mode = selected_ai == "devin"
+        native_skill_mode = codex_skill_mode or claude_skill_mode or kimi_skill_mode or agy_skill_mode or trae_skill_mode or cursor_agent_skill_mode or copilot_skill_mode or devin_skill_mode
+
+        if codex_skill_mode and not ai_skills:
+            steps_lines.append(f"{step_num}. Start Codex in this project directory; spec-kit skills were installed to [cyan].agents/skills[/cyan]")
+            step_num += 1
+        if claude_skill_mode and not ai_skills:
+            steps_lines.append(f"{step_num}. Start Claude in this project directory; spec-kit skills were installed to [cyan].claude/skills[/cyan]")
+            step_num += 1
+        if cursor_agent_skill_mode and not ai_skills:
+            steps_lines.append(f"{step_num}. Start Cursor Agent in this project directory; spec-kit skills were installed to [cyan].cursor/skills[/cyan]")
+            step_num += 1
+        if devin_skill_mode:
+            steps_lines.append(f"{step_num}. Start Devin in this project directory; spec-kit skills were installed to [cyan].devin/skills[/cyan]")
+            step_num += 1
+        usage_label = "skills" if native_skill_mode else "slash commands"
+
+        def _display_cmd(name: str) -> str:
+            if codex_skill_mode or agy_skill_mode or trae_skill_mode:
+                return f"$speckit-{name}"
+            if claude_skill_mode:
+                return f"/speckit-{name}"
+            if kimi_skill_mode:
+                return f"/skill:speckit-{name}"
+            if cursor_agent_skill_mode or copilot_skill_mode or devin_skill_mode:
+                return f"/speckit-{name}"
+            return f"/speckit.{name}"
+
+        steps_lines.append(f"{step_num}. Start using {usage_label} with your coding agent:")
+
+        steps_lines.append(f"   {step_num}.1 [cyan]{_display_cmd('constitution')}[/] - Establish project principles")
+        steps_lines.append(f"   {step_num}.2 [cyan]{_display_cmd('specify')}[/] - Create baseline specification")
+        steps_lines.append(f"   {step_num}.3 [cyan]{_display_cmd('plan')}[/] - Create implementation plan")
+        steps_lines.append(f"   {step_num}.4 [cyan]{_display_cmd('tasks')}[/] - Generate actionable tasks")
+        steps_lines.append(f"   {step_num}.5 [cyan]{_display_cmd('implement')}[/] - Execute implementation")
+
+        steps_panel = Panel("\n".join(steps_lines), title="Next Steps", border_style="cyan", padding=(1, 2))
+        console.print()
+        console.print(steps_panel)
+
+        enhancement_intro = (
+            "Optional skills that you can use for your specs [bright_black](improve quality & confidence)[/bright_black]"
+            if native_skill_mode
+            else "Optional commands that you can use for your specs [bright_black](improve quality & confidence)[/bright_black]"
+        )
+        enhancement_lines = [
+            enhancement_intro,
+            "",
+            f"○ [cyan]{_display_cmd('clarify')}[/] [bright_black](optional)[/bright_black] - Ask structured questions to de-risk ambiguous areas before planning (run before [cyan]{_display_cmd('plan')}[/] if used)",
+            f"○ [cyan]{_display_cmd('analyze')}[/] [bright_black](optional)[/bright_black] - Cross-artifact consistency & alignment report (after [cyan]{_display_cmd('tasks')}[/], before [cyan]{_display_cmd('implement')}[/])",
+            f"○ [cyan]{_display_cmd('checklist')}[/] [bright_black](optional)[/bright_black] - Generate quality checklists to validate requirements completeness, clarity, and consistency (after [cyan]{_display_cmd('plan')}[/])"
+        ]
+        enhancements_title = "Enhancement Skills" if native_skill_mode else "Enhancement Commands"
+        enhancements_panel = Panel("\n".join(enhancement_lines), title=enhancements_title, border_style="cyan", padding=(1, 2))
+        console.print()
+        console.print(enhancements_panel)

src/specify_cli/commands/integration.py

@@ -0,0 +1,2 @@
+"""specify integration * commands — placeholder for future extraction."""
+from __future__ import annotations

src/specify_cli/commands/preset.py

@@ -0,0 +1,2 @@
+"""specify preset * commands — placeholder for future extraction."""
+from __future__ import annotations

src/specify_cli/commands/workflow.py

@@ -0,0 +1,2 @@
+"""specify workflow * commands — placeholder for future extraction."""
+from __future__ import annotations

src/specify_cli/extensions.py

@@ -801,42 +801,29 @@ class ExtensionManager:
     def _get_skills_dir(self) -> Optional[Path]:
         """Return the active skills directory for extension skill registration.
 
-        Reads ``.specify/init-options.json`` to determine whether skills
-        are enabled and which agent was selected, then delegates to
-        the module-level ``_get_skills_dir()`` helper for the concrete path.
+        Delegates to :func:`resolve_active_skills_dir` which reads
+        init-options, applies the Kimi native-skills fallback, and
+        safely creates the directory when ``ai_skills`` is enabled.
 
-        Kimi is treated as a native-skills agent: if ``ai == "kimi"`` and
-        ``.kimi/skills`` exists, extension installs should still propagate
-        command skills even when ``ai_skills`` is false.
-
-        Returns:
-            The skills directory ``Path``, or ``None`` if skills were not
-            enabled and no native-skills fallback applies.
+        Returns ``None`` (instead of raising) when the directory cannot
+        be created due to symlink, containment, or permission issues so
+        that callers can fall back gracefully.
         """
-        from . import load_init_options, _get_skills_dir as resolve_skills_dir
-
-        opts = load_init_options(self.project_root)
-        if not isinstance(opts, dict):
-            opts = {}
-
-        agent = opts.get("ai")
-        if not isinstance(agent, str) or not agent:
+        from . import resolve_active_skills_dir, _print_cli_warning
+        try:
+            return resolve_active_skills_dir(self.project_root)
+        except (ValueError, OSError) as exc:
+            _print_cli_warning(
+                "resolve", "skills directory", None, exc,
+                continuing="Continuing without skill registration.",
+            )
             return None
 
-        ai_skills_enabled = bool(opts.get("ai_skills"))
-        if not ai_skills_enabled and agent != "kimi":
-            return None
-
-        skills_dir = resolve_skills_dir(self.project_root, agent)
-        if not skills_dir.is_dir():
-            return None
-
-        return skills_dir
-
     def _register_extension_skills(
         self,
         manifest: ExtensionManifest,
         extension_dir: Path,
+        link_outputs: bool = False,
     ) -> List[str]:
         """Generate SKILL.md files for extension commands as agent skills.
 
@@ -848,6 +835,8 @@ class ExtensionManager:
         Args:
             manifest: Extension manifest.
             extension_dir: Installed extension directory.
+            link_outputs: If True, create dev-mode symlinks for rendered
+                skill files when supported by the OS.
 
         Returns:
             List of skill names that were created (for registry storage).
@@ -900,9 +889,18 @@ class ExtensionManager:
             # Check if skill already exists before creating the directory
             skill_subdir = skills_dir / skill_name
             skill_file = skill_subdir / "SKILL.md"
-            if skill_file.exists():
-                # Do not overwrite user-customized skills
-                continue
+            cache_root = extension_dir / ".specify-dev" / "extension-skills"
+            cache_file = cache_root / skill_name / "SKILL.md"
+            CommandRegistrar._ensure_inside(cache_file, cache_root)
+            if skill_file.exists() or skill_file.is_symlink():
+                # Do not overwrite user-customized skills, but allow dev-mode
+                # symlinks that point back to this extension's generated cache
+                # to be refreshed on a subsequent dev install.
+                if not (
+                    link_outputs
+                    and self._is_expected_dev_symlink(skill_file, cache_file)
+                ):
+                    continue
 
             # Create skill directory; track whether we created it so we can clean
             # up safely if reading the source file subsequently fails.
@@ -954,11 +952,35 @@ class ExtensionManager:
                     skill_content
                 )
 
-            skill_file.write_text(skill_content, encoding="utf-8")
+            if link_outputs:
+                try:
+                    cache_file.parent.mkdir(parents=True, exist_ok=True)
+                    cache_file.write_text(skill_content, encoding="utf-8")
+                    if skill_file.exists() or skill_file.is_symlink():
+                        skill_file.unlink()
+                    target = os.path.relpath(cache_file, skill_file.parent)
+                    os.symlink(target, skill_file)
+                except (OSError, ValueError):
+                    if skill_file.is_symlink():
+                        skill_file.unlink()
+                    skill_file.write_text(skill_content, encoding="utf-8")
+            else:
+                skill_file.write_text(skill_content, encoding="utf-8")
             written.append(skill_name)
 
         return written
 
+    @staticmethod
+    def _is_expected_dev_symlink(skill_file: Path, cache_file: Path) -> bool:
+        """Return True when an existing skill file links to its dev cache."""
+        if not skill_file.is_symlink():
+            return False
+
+        try:
+            return skill_file.resolve(strict=False) == cache_file.resolve(strict=False)
+        except OSError:
+            return False
+
     def _unregister_extension_skills(
         self,
         skill_names: List[str],
@@ -1129,6 +1151,7 @@ class ExtensionManager:
         speckit_version: str,
         register_commands: bool = True,
         priority: int = 10,
+        link_commands: bool = False,
     ) -> ExtensionManifest:
         """Install extension from a local directory.
 
@@ -1137,6 +1160,8 @@ class ExtensionManager:
             speckit_version: Current spec-kit version
             register_commands: If True, register commands with AI agents
             priority: Resolution priority (lower = higher precedence, default 10)
+            link_commands: If True, register rendered agent artifacts as
+                symlinks to a dev cache when supported by the OS.
 
         Returns:
             Installed extension manifest
@@ -1180,12 +1205,14 @@ class ExtensionManager:
             registrar = CommandRegistrar()
             # Register for all detected agents
             registered_commands = registrar.register_commands_for_all_agents(
-                manifest, dest_dir, self.project_root
+                manifest, dest_dir, self.project_root, link_outputs=link_commands
             )
 
         # Auto-register extension commands as agent skills when --ai-skills
         # was used during project initialisation (feature parity).
-        registered_skills = self._register_extension_skills(manifest, dest_dir)
+        registered_skills = self._register_extension_skills(
+            manifest, dest_dir, link_outputs=link_commands
+        )
 
         # Register hooks and update installed list in extensions.yml
         hook_executor = HookExecutor(self.project_root)
@@ -1621,7 +1648,8 @@ class CommandRegistrar:
         agent_name: str,
         manifest: ExtensionManifest,
         extension_dir: Path,
-        project_root: Path
+        project_root: Path,
+        link_outputs: bool = False,
     ) -> List[str]:
         """Register extension commands for a specific agent."""
         if agent_name not in self.AGENT_CONFIGS:
@@ -1629,20 +1657,23 @@ class CommandRegistrar:
         context_note = f"\n<!-- Extension: {manifest.id} -->\n<!-- Config: .specify/extensions/{manifest.id}/ -->\n"
         return self._registrar.register_commands(
             agent_name, manifest.commands, manifest.id, extension_dir, project_root,
-            context_note=context_note
+            context_note=context_note,
+            link_outputs=link_outputs,
         )
 
     def register_commands_for_all_agents(
         self,
         manifest: ExtensionManifest,
         extension_dir: Path,
-        project_root: Path
+        project_root: Path,
+        link_outputs: bool = False,
     ) -> Dict[str, List[str]]:
         """Register extension commands for all detected agents."""
         context_note = f"\n<!-- Extension: {manifest.id} -->\n<!-- Config: .specify/extensions/{manifest.id}/ -->\n"
         return self._registrar.register_commands_for_all_agents(
             manifest.commands, manifest.id, extension_dir, project_root,
-            context_note=context_note
+            context_note=context_note,
+            link_outputs=link_outputs,
         )
 
     def unregister_commands(
@@ -1657,10 +1688,13 @@ class CommandRegistrar:
         self,
         manifest: ExtensionManifest,
         extension_dir: Path,
-        project_root: Path
+        project_root: Path,
+        link_outputs: bool = False,
     ) -> List[str]:
         """Register extension commands for Claude Code agent."""
-        return self.register_commands_for_agent("claude", manifest, extension_dir, project_root)
+        return self.register_commands_for_agent(
+            "claude", manifest, extension_dir, project_root, link_outputs=link_outputs
+        )
 
 
 class ExtensionCatalog(CatalogStackBase):

src/specify_cli/integrations/__init__.py

@@ -61,6 +61,7 @@ def _register_builtins() -> None:
     from .gemini import GeminiIntegration
     from .generic import GenericIntegration
     from .goose import GooseIntegration
+    from .hermes import HermesIntegration
     from .iflow import IflowIntegration
     from .junie import JunieIntegration
     from .kilocode import KilocodeIntegration
@@ -93,6 +94,7 @@ def _register_builtins() -> None:
     _register(GeminiIntegration())
     _register(GenericIntegration())
     _register(GooseIntegration())
+    _register(HermesIntegration())
     _register(IflowIntegration())
     _register(JunieIntegration())
     _register(KilocodeIntegration())

src/specify_cli/integrations/agy/__init__.py

@@ -5,6 +5,7 @@ Antigravity uses ``.agents/skills/speckit-<name>/SKILL.md`` layout (enforced sin
 
 from __future__ import annotations
 
+import re
 from pathlib import Path
 from typing import TYPE_CHECKING, Any
 
@@ -13,6 +14,15 @@ from ..base import SkillsIntegration
 if TYPE_CHECKING:
     from ..manifest import IntegrationManifest
 
+# Note injected into hook sections so agy maps dot-notation command
+# names (from extensions.yml) to the hyphenated skill names it uses.
+# Without this, agy emits ``/speckit.git.commit`` (which does not
+# resolve) instead of ``/speckit-git-commit``.
+_HOOK_COMMAND_NOTE = (
+    "- When constructing slash commands from hook command names, "
+    "replace dots (`.`) with hyphens (`-`). "
+    "For example, `speckit.git.commit` → `/speckit-git-commit`.\n"
+)
 
 
 class AgyIntegration(SkillsIntegration):
@@ -23,8 +33,8 @@ class AgyIntegration(SkillsIntegration):
         "name": "Antigravity",
         "folder": ".agents/",
         "commands_subdir": "skills",
-        "install_url": None,
-        "requires_cli": False,
+        "install_url": "https://antigravity.google/",
+        "requires_cli": True,
     }
     registrar_config = {
         "dir": ".agents/skills",
@@ -34,6 +44,54 @@ class AgyIntegration(SkillsIntegration):
     }
     context_file = "AGENTS.md"
 
+    @staticmethod
+    def _inject_hook_command_note(content: str) -> str:
+        """Insert a dot-to-hyphen note before each hook output instruction.
+
+        Targets the line ``- For each executable hook, output the following``
+        and inserts the note on the line before it, matching its indentation.
+        Skips if the note is already present.
+        """
+        if "replace dots" in content:
+            return content
+
+        def repl(m: re.Match[str]) -> str:
+            indent = m.group(1)
+            instruction = m.group(2)
+            # ``eol`` is empty when the regex matched via ``$`` because the
+            # instruction was the final line of a file with no trailing
+            # newline. Default to ``\n`` so the note never collapses onto
+            # the same line as the instruction.
+            eol = m.group(3) or "\n"
+            return (
+                indent
+                + _HOOK_COMMAND_NOTE.rstrip("\n")
+                + eol
+                + indent
+                + instruction
+                + eol
+            )
+
+        return re.sub(
+            r"(?m)^(\s*)(- For each executable hook, output the following[^\r\n]*)(\r\n|\n|$)",
+            repl,
+            content,
+        )
+
+    def post_process_skill_content(self, content: str) -> str:
+        """Inject the dot-to-hyphen hook command note."""
+        return self._inject_hook_command_note(content)
+
+    def build_exec_args(
+        self,
+        prompt: str,
+        *,
+        model: str | None = None,
+        output_json: bool = True,
+    ) -> list[str] | None:
+        # agy does not support --model or JSON output; both params are ignored
+        return [self._resolve_executable(), "--print", prompt]
+
     def setup(
         self,
         project_root: Path,
@@ -49,4 +107,21 @@ class AgyIntegration(SkillsIntegration):
             fg="yellow",
             err=True,
         )
-        return super().setup(project_root, manifest, parsed_options=parsed_options, **opts)
+        created = super().setup(project_root, manifest, parsed_options=parsed_options, **opts)
+
+        skills_dir = self.skills_dest(project_root).resolve()
+        for path in created:
+            try:
+                path.resolve().relative_to(skills_dir)
+            except ValueError:
+                continue
+            if path.name != "SKILL.md":
+                continue
+
+            content = path.read_bytes().decode("utf-8")
+            updated = self.post_process_skill_content(content)
+            if updated != content:
+                path.write_bytes(updated.encode("utf-8"))
+                self.record_file_in_manifest(path, project_root, manifest)
+
+        return created

src/specify_cli/integrations/base.py

@@ -13,7 +13,9 @@ Provides:
 
 from __future__ import annotations
 
+import os
 import re
+import shlex
 import shutil
 from abc import ABC
 from dataclasses import dataclass
@@ -25,6 +27,12 @@ import yaml
 if TYPE_CHECKING:
     from .manifest import IntegrationManifest
 
+_HOOK_COMMAND_NOTE = (
+    "- When constructing slash commands from hook command names, "
+    "replace dots (`.`) with hyphens (`-`). "
+    "For example, `speckit.git.commit` → `/speckit-git-commit`.\n"
+)
+
 
 # ---------------------------------------------------------------------------
 # IntegrationOption
@@ -138,6 +146,65 @@ class IntegrationBase(ABC):
         """
         return None
 
+    def _resolve_executable(self) -> str:
+        """Return the executable for this integration's CLI tool.
+
+        Checks ``SPECKIT_INTEGRATION_<KEY>_EXECUTABLE`` first, allowing
+        operators to override the binary path without modifying the
+        integration configuration — useful when the tool is installed in
+        a non-standard location or a specific version must be pinned.
+        Hyphens in the integration key are replaced with underscores and
+        the key is uppercased so that, for example, ``kiro-cli`` maps to
+        ``SPECKIT_INTEGRATION_KIRO_CLI_EXECUTABLE``.
+
+        Falls back to ``self.key`` when the env var is unset or
+        whitespace-only so existing behaviour is unchanged.
+
+        See issue #2596.
+        """
+        env_name = (
+            f"SPECKIT_INTEGRATION_{self.key.upper().replace('-', '_')}_EXECUTABLE"
+        )
+        override = os.environ.get(env_name, "").strip()
+        return override if override else self.key
+
+    def _apply_extra_args_env_var(self, args: list[str]) -> None:
+        """Append `SPECKIT_INTEGRATION_<KEY>_EXTRA_ARGS` env-var value to *args*.
+
+        Operators can inject extra CLI flags into the spawned agent
+        subprocess by setting an env var named for the integration key,
+        e.g. `SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS="--dangerously-skip-permissions"`.
+        The `INTEGRATION` segment scopes the variable to this subsystem
+        so it does not collide with other Spec Kit env-var namespaces.
+        Hyphens in the integration key are replaced with underscores
+        and the key is uppercased
+        (e.g. `kiro-cli` → `SPECKIT_INTEGRATION_KIRO_CLI_EXTRA_ARGS`).
+
+        Useful in CI / non-interactive contexts where the spawned agent
+        needs flags that change its prompt-handling behaviour.
+        Default behaviour (env var unset or whitespace-only) is a no-op
+        — *args* is unchanged. Multi-token values are parsed via
+        `shlex.split`.
+
+        See issue #2595.
+        """
+        env_name = (
+            f"SPECKIT_INTEGRATION_{self.key.upper().replace('-', '_')}_EXTRA_ARGS"
+        )
+        extra = os.environ.get(env_name, "").strip()
+        if not extra:
+            return
+        try:
+            tokens = shlex.split(extra)
+        except ValueError as exc:
+            raise ValueError(
+                f"{env_name} is not parseable as a POSIX-quoted command line "
+                f"(value: {extra!r}). shlex reported: {exc}. "
+                f"Use single or double quotes to group multi-word values, e.g. "
+                f'{env_name}=\'--flag "value with spaces"\'.'
+            ) from exc
+        args.extend(tokens)
+
     def build_command_invocation(self, command_name: str, args: str = "") -> str:
         """Build the native slash-command invocation for a Spec Kit command.
 
@@ -850,7 +917,8 @@ class MarkdownIntegration(IntegrationBase):
     ) -> list[str] | None:
         if not self.config or not self.config.get("requires_cli"):
             return None
-        args = [self.key, "-p", prompt]
+        args = [self._resolve_executable(), "-p", prompt]
+        self._apply_extra_args_env_var(args)
         if model:
             args.extend(["--model", model])
         if output_json:
@@ -937,7 +1005,8 @@ class TomlIntegration(IntegrationBase):
     ) -> list[str] | None:
         if not self.config or not self.config.get("requires_cli"):
             return None
-        args = [self.key, "-p", prompt]
+        args = [self._resolve_executable(), "-p", prompt]
+        self._apply_extra_args_env_var(args)
         if model:
             args.extend(["-m", model])
         if output_json:
@@ -1355,7 +1424,8 @@ class SkillsIntegration(IntegrationBase):
     ) -> list[str] | None:
         if not self.config or not self.config.get("requires_cli"):
             return None
-        args = [self.key, "-p", prompt]
+        args = [self._resolve_executable(), "-p", prompt]
+        self._apply_extra_args_env_var(args)
         if model:
             args.extend(["--model", model])
         if output_json:
@@ -1391,15 +1461,53 @@ class SkillsIntegration(IntegrationBase):
             invocation = f"{invocation} {args}"
         return invocation
 
+    @staticmethod
+    def _inject_hook_command_note(content: str) -> str:
+        """Insert a dot-to-hyphen note before each hook output instruction.
+
+        Targets the line ``- For each executable hook, output the following``
+        and inserts the note on the line before it, matching its indentation.
+        Skips individual instructions that already have the note immediately
+        above them.
+        """
+        note = _HOOK_COMMAND_NOTE.rstrip("\n")
+
+        def repl(m: re.Match[str]) -> str:
+            indent = m.group(1)
+            instruction = m.group(2)
+            previous_lines = content[:m.start()].splitlines()
+            if previous_lines and previous_lines[-1] == indent + note:
+                return m.group(0)
+            # ``eol`` is empty when the regex matched via ``$`` because the
+            # instruction was the final line of a file with no trailing
+            # newline. Default to ``\n`` so the note never collapses onto
+            # the same line as the instruction.
+            eol = m.group(3) or "\n"
+            return (
+                indent
+                + note
+                + eol
+                + indent
+                + instruction
+                + eol
+            )
+
+        return re.sub(
+            r"(?m)^([ \t]*)(- For each executable hook, output the following[^\r\n]*)(\r\n|\n|$)",
+            repl,
+            content,
+        )
+
     def post_process_skill_content(self, content: str) -> str:
         """Post-process a SKILL.md file's content after generation.
 
         Called by external skill generators (presets, extensions) to let
         the integration inject agent-specific frontmatter or body
-        transformations.  The default implementation returns *content*
-        unchanged.  Subclasses may override — see ``ClaudeIntegration``.
+        transformations.  The base implementation injects shared skills
+        guidance for converting dotted hook command names to hyphenated
+        slash commands.  Subclasses may override — see ``ClaudeIntegration``.
         """
-        return content
+        return self._inject_hook_command_note(content)
 
     def setup(
         self,
@@ -1502,6 +1610,8 @@ class SkillsIntegration(IntegrationBase):
                 f"{processed_body}"
             )
 
+            skill_content = self.post_process_skill_content(skill_content)
+
             # Write speckit-<name>/SKILL.md
             skill_dir = skills_dir / skill_name
             skill_file = skill_dir / "SKILL.md"

src/specify_cli/integrations/claude/__init__.py

@@ -5,21 +5,11 @@ from __future__ import annotations
 from pathlib import Path
 from typing import Any
 
-import re
-
 import yaml
 
 from ..base import SkillsIntegration
 from ..manifest import IntegrationManifest
 
-# Note injected into hook sections so Claude maps dot-notation command
-# names (from extensions.yml) to the hyphenated skill names it uses.
-_HOOK_COMMAND_NOTE = (
-    "- When constructing slash commands from hook command names, "
-    "replace dots (`.`) with hyphens (`-`). "
-    "For example, `speckit.git.commit` → `/speckit-git-commit`.\n"
-)
-
 # Mapping of command template stem → argument-hint text shown inline
 # when a user invokes the slash command in Claude Code.
 ARGUMENT_HINTS: dict[str, str] = {
@@ -159,41 +149,11 @@ class ClaudeIntegration(SkillsIntegration):
             out.append(line)
         return "".join(out)
 
-    @staticmethod
-    def _inject_hook_command_note(content: str) -> str:
-        """Insert a dot-to-hyphen note before each hook output instruction.
-
-        Targets the line ``- For each executable hook, output the following``
-        and inserts the note on the line before it, matching its indentation.
-        Skips if the note is already present.
-        """
-        if "replace dots" in content:
-            return content
-
-        def repl(m: re.Match[str]) -> str:
-            indent = m.group(1)
-            instruction = m.group(2)
-            eol = m.group(3)
-            return (
-                indent
-                + _HOOK_COMMAND_NOTE.rstrip("\n")
-                + eol
-                + indent
-                + instruction
-                + eol
-            )
-
-        return re.sub(
-            r"(?m)^(\s*)(- For each executable hook, output the following[^\r\n]*)(\r\n|\n|$)",
-            repl,
-            content,
-        )
-
     def post_process_skill_content(self, content: str) -> str:
         """Inject Claude-specific frontmatter flags and hook notes."""
-        updated = self._inject_frontmatter_flag(content, "user-invocable")
+        updated = super().post_process_skill_content(content)
+        updated = self._inject_frontmatter_flag(updated, "user-invocable")
         updated = self._inject_frontmatter_flag(updated, "disable-model-invocation", "false")
-        updated = self._inject_hook_command_note(updated)
         return updated
 
     def setup(
@@ -203,10 +163,9 @@ class ClaudeIntegration(SkillsIntegration):
         parsed_options: dict[str, Any] | None = None,
         **opts: Any,
     ) -> list[Path]:
-        """Install Claude skills, then inject Claude-specific flags and argument-hints."""
+        """Install Claude skills, then inject argument-hints."""
         created = super().setup(project_root, manifest, parsed_options, **opts)
 
-        # Post-process generated skill files
         skills_dir = self.skills_dest(project_root).resolve()
 
         for path in created:
@@ -221,7 +180,7 @@ class ClaudeIntegration(SkillsIntegration):
             content_bytes = path.read_bytes()
             content = content_bytes.decode("utf-8")
 
-            updated = self.post_process_skill_content(content)
+            updated = content
 
             # Inject argument-hint if available for this skill
             skill_dir_name = path.parent.name  # e.g. "speckit-plan"

src/specify_cli/integrations/codex/__init__.py

@@ -6,22 +6,7 @@ Commands are deprecated; ``--skills`` defaults to ``True``.
 
 from __future__ import annotations
 
-import re
-from pathlib import Path
-from typing import Any
-
 from ..base import IntegrationOption, SkillsIntegration
-from ..manifest import IntegrationManifest
-
-# Note injected into hook sections so Codex maps dot-notation command
-# names (from extensions.yml) to the hyphenated skill names it uses.
-# Without this, Codex emits ``/speckit.git.commit`` (which does not
-# resolve) instead of ``/speckit-git-commit``.
-_HOOK_COMMAND_NOTE = (
-    "- When constructing slash commands from hook command names, "
-    "replace dots (`.`) with hyphens (`-`). "
-    "For example, `speckit.git.commit` → `/speckit-git-commit`.\n"
-)
 
 
 class CodexIntegration(SkillsIntegration):
@@ -52,7 +37,10 @@ class CodexIntegration(SkillsIntegration):
         output_json: bool = True,
     ) -> list[str] | None:
         # Codex uses ``codex exec "prompt"`` for non-interactive mode.
-        args: list[str] = ["codex", "exec", prompt]
+        # Resolve argv[0] via the shared executable resolver so operators can
+        # override the binary with SPECKIT_INTEGRATION_CODEX_EXECUTABLE.
+        args: list[str] = [self._resolve_executable(), "exec", prompt]
+        self._apply_extra_args_env_var(args)
         if model:
             args.extend(["--model", model])
         if output_json:
@@ -69,68 +57,3 @@ class CodexIntegration(SkillsIntegration):
                 help="Install as agent skills (default for Codex)",
             ),
         ]
-
-    @staticmethod
-    def _inject_hook_command_note(content: str) -> str:
-        """Insert a dot-to-hyphen note before each hook output instruction.
-
-        Targets the line ``- For each executable hook, output the following``
-        and inserts the note on the line before it, matching its indentation.
-        Skips if the note is already present.
-        """
-        if "replace dots" in content:
-            return content
-
-        def repl(m: re.Match[str]) -> str:
-            indent = m.group(1)
-            instruction = m.group(2)
-            # ``eol`` is empty when the regex matched via ``$`` because the
-            # instruction was the final line of a file with no trailing
-            # newline. Default to ``\n`` so the note never collapses onto
-            # the same line as the instruction.
-            eol = m.group(3) or "\n"
-            return (
-                indent
-                + _HOOK_COMMAND_NOTE.rstrip("\n")
-                + eol
-                + indent
-                + instruction
-                + eol
-            )
-
-        return re.sub(
-            r"(?m)^(\s*)(- For each executable hook, output the following[^\r\n]*)(\r\n|\n|$)",
-            repl,
-            content,
-        )
-
-    def post_process_skill_content(self, content: str) -> str:
-        """Inject the dot-to-hyphen hook command note."""
-        return self._inject_hook_command_note(content)
-
-    def setup(
-        self,
-        project_root: Path,
-        manifest: IntegrationManifest,
-        parsed_options: dict[str, Any] | None = None,
-        **opts: Any,
-    ) -> list[Path]:
-        """Install Codex skills, then inject the hook command note."""
-        created = super().setup(project_root, manifest, parsed_options, **opts)
-
-        skills_dir = self.skills_dest(project_root).resolve()
-        for path in created:
-            try:
-                path.resolve().relative_to(skills_dir)
-            except ValueError:
-                continue
-            if path.name != "SKILL.md":
-                continue
-
-            content = path.read_bytes().decode("utf-8")
-            updated = self.post_process_skill_content(content)
-            if updated != content:
-                path.write_bytes(updated.encode("utf-8"))
-                self.record_file_in_manifest(path, project_root, manifest)
-
-        return created

src/specify_cli/integrations/copilot/__init__.py

@@ -24,6 +24,16 @@ from ..base import IntegrationBase, IntegrationOption, SkillsIntegration
 from ..manifest import IntegrationManifest
 
 
+def _copilot_executable() -> str:
+    """Return the executable name for Copilot CLI on this platform.
+
+    On Windows, subprocess invocation is reliable with `copilot.cmd`.
+    """
+    if os.name == "nt":
+        return "copilot.cmd"
+    return "copilot"
+
+
 def _allow_all() -> bool:
     """Return True if the Copilot CLI should run with full permissions.
 
@@ -124,6 +134,18 @@ class CopilotIntegration(IntegrationBase):
             ),
         ]
 
+    def _resolve_executable(self) -> str:
+        """Return the Copilot CLI executable, respecting the env-var override.
+
+        Checks ``SPECKIT_INTEGRATION_COPILOT_EXECUTABLE`` first.  Falls
+        back to the platform-specific default from ``_copilot_executable()``
+        (``copilot.cmd`` on Windows, ``copilot`` elsewhere) so that
+        existing behaviour is preserved when the env var is unset.
+        """
+        env_name = "SPECKIT_INTEGRATION_COPILOT_EXECUTABLE"
+        override = os.environ.get(env_name, "").strip()
+        return override if override else _copilot_executable()
+
     def build_exec_args(
         self,
         prompt: str,
@@ -138,7 +160,8 @@ class CopilotIntegration(IntegrationBase):
         # Controlled by SPECKIT_COPILOT_ALLOW_ALL_TOOLS env var
         # (default: enabled).  The deprecated SPECKIT_ALLOW_ALL_TOOLS
         # is also honoured as a fallback.
-        args = ["copilot", "-p", prompt]
+        args = [self._resolve_executable(), "-p", prompt]
+        self._apply_extra_args_env_var(args)
         if _allow_all():
             args.append("--yolo")
         if model:
@@ -206,7 +229,12 @@ class CopilotIntegration(IntegrationBase):
             agent_name = f"speckit.{stem}"
             prompt = args or ""
 
-        cli_args = ["copilot", "-p", prompt]
+        cli_args = [self._resolve_executable(), "-p", prompt]
+        # Honour SPECKIT_INTEGRATION_COPILOT_EXTRA_ARGS for real workflow
+        # runs.  `dispatch_command` builds cli_args inline rather than
+        # going through `build_exec_args`, so the hook must be invoked
+        # here too — otherwise the env var is silently ignored.
+        self._apply_extra_args_env_var(cli_args)
         if not skills_mode:
             cli_args.extend(["--agent", agent_name])
         if _allow_all():
@@ -255,12 +283,13 @@ class CopilotIntegration(IntegrationBase):
         return f"speckit.{template_name}.agent.md"
 
     def post_process_skill_content(self, content: str) -> str:
-        """Inject Copilot-specific ``mode:`` field into SKILL.md frontmatter.
+        """Inject shared hook guidance and Copilot ``mode:`` frontmatter.
 
         Inserts ``mode: speckit.<stem>`` before the closing ``---`` so
         Copilot can associate the skill with its agent mode.
         """
-        lines = content.splitlines(keepends=True)
+        updated = _CopilotSkillsHelper().post_process_skill_content(content)
+        lines = updated.splitlines(keepends=True)
 
         # Extract skill name from frontmatter to derive the mode value
         dash_count = 0
@@ -274,7 +303,7 @@ class CopilotIntegration(IntegrationBase):
                 continue
             if dash_count == 1:
                 if stripped.startswith("mode:"):
-                    return content  # already present
+                    return updated  # already present
                 if stripped.startswith("name:"):
                     # Parse: name: "speckit-plan" → speckit.plan
                     val = stripped.split(":", 1)[1].strip().strip('"').strip("'")
@@ -285,7 +314,7 @@ class CopilotIntegration(IntegrationBase):
                         skill_name = val
 
         if not skill_name:
-            return content
+            return updated
 
         # Inject mode: before the closing --- of frontmatter
         out: list[str] = []

src/specify_cli/integrations/devin/__init__.py

@@ -48,7 +48,8 @@ class DevinIntegration(SkillsIntegration):
         stdout instead of structured JSON. ``requires_cli=True`` is
         kept on the integration for tool detection.
         """
-        args = [self.key, "-p", prompt]
+        args = [self._resolve_executable(), "-p", prompt]
+        self._apply_extra_args_env_var(args)
         if model:
             args.extend(["--model", model])
         return args

src/specify_cli/integrations/hermes/__init__.py

@@ -0,0 +1,280 @@
+"""Hermes Agent integration — skills-based agent.
+
+Hermes Agent (https://github.com/NousResearch/hermes-agent) is an open-source
+AI agent framework by Nous Research.  It stores skills in
+``~/.hermes/skills/`` (user-global) rather than a project-local directory.
+
+Usage::
+
+    specify init my-project --integration hermes
+    specify init --here --ai hermes
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+from shutil import rmtree
+from typing import Any
+
+import yaml
+
+from ..base import IntegrationOption, SkillsIntegration
+from ..manifest import IntegrationManifest
+
+
+class HermesIntegration(SkillsIntegration):
+    """Integration for Hermes Agent skills.
+
+    Hermes loads skills from ``~/.hermes/skills/`` (user home directory)
+    rather than a project-local path.  Skills are installed directly to
+    the global directory — no project-local copies are created since
+    Hermes discovers them globally.  A project-local marker directory
+    (``.hermes/skills/`` empty) is created so extension commands (e.g.
+    git) can detect Hermes as an active integration.  Uninstall removes
+    both the marker and all global ``speckit-*`` skills, matching the
+    standard integration teardown behaviour.
+    """
+
+    key = "hermes"
+    config = {
+        "name": "Hermes Agent",
+        "folder": ".hermes/",
+        "commands_subdir": "skills",
+        "install_url": "https://github.com/NousResearch/hermes-agent",
+        "requires_cli": True,
+    }
+    registrar_config = {
+        "dir": "~/.hermes/skills",
+        "detect_dir": ".hermes/skills",
+        "format": "markdown",
+        "args": "$ARGUMENTS",
+        "extension": "/SKILL.md",
+    }
+    context_file = "AGENTS.md"
+
+    # -- Helpers -----------------------------------------------------------
+
+    @staticmethod
+    def _hermes_home_skills_dir() -> Path:
+        """Return ``~/.hermes/skills/`` — the global skills directory."""
+        return Path.home() / ".hermes" / "skills"
+
+    # -- Options -----------------------------------------------------------
+
+    @classmethod
+    def options(cls) -> list[IntegrationOption]:
+        return [
+            IntegrationOption(
+                "--skills",
+                is_flag=True,
+                default=True,
+                help="Install as agent skills (default for Hermes Agent)",
+            ),
+        ]
+
+    # -- Setup -------------------------------------------------------------
+
+    def setup(
+        self,
+        project_root: Path,
+        manifest: IntegrationManifest,
+        parsed_options: dict[str, Any] | None = None,
+        **opts: Any,
+    ) -> list[Path]:
+        """Install command templates as global Hermes skills.
+
+        Writes each skill directly to
+        ``~/.hermes/skills/speckit-<name>/SKILL.md`` where Hermes
+        discovers them at runtime.  No project-local SKILL.md copies are
+        created — the global directory is the single source of truth.
+        A project-local marker (``.hermes/skills/`` empty) is created
+        so extension commands (e.g. git) can detect Hermes as an active
+        integration.
+        """
+        templates = self.list_command_templates()
+        if not templates:
+            return []
+
+        # Safety check: verify manifest project_root matches (standard pattern)
+        project_root_resolved = project_root.resolve()
+        if manifest.project_root != project_root_resolved:
+            raise ValueError(
+                f"manifest.project_root ({manifest.project_root}) does not match "
+                f"project_root ({project_root_resolved})"
+            )
+
+        script_type = opts.get("script_type", "sh")
+        arg_placeholder = (
+            self.registrar_config.get("args", "$ARGUMENTS")
+            if self.registrar_config
+            else "$ARGUMENTS"
+        )
+
+        global_skills_dir = self._hermes_home_skills_dir()
+        global_skills_dir.mkdir(parents=True, exist_ok=True)
+
+        created: list[Path] = []
+
+        for src_file in templates:
+            raw = src_file.read_text(encoding="utf-8")
+
+            # Derive the skill name from the template stem
+            command_name = src_file.stem  # e.g. "plan"
+            skill_name = f"speckit-{command_name.replace('.', '-')}"
+
+            # Parse frontmatter for description
+            frontmatter: dict[str, Any] = {}
+            if raw.startswith("---"):
+                parts = raw.split("---", 2)
+                if len(parts) >= 3:
+                    try:
+                        fm = yaml.safe_load(parts[1])
+                        if isinstance(fm, dict):
+                            frontmatter = fm
+                    except yaml.YAMLError:
+                        pass
+
+            # Process body through the standard template pipeline
+            processed_body = self.process_template(
+                raw,
+                self.key,
+                script_type,
+                arg_placeholder,
+                context_file=self.context_file or "",
+                invoke_separator=self.invoke_separator,
+            )
+            # Strip the processed frontmatter — we rebuild it for skills.
+            if processed_body.startswith("---"):
+                parts = processed_body.split("---", 2)
+                if len(parts) >= 3:
+                    processed_body = parts[2]
+
+            # Select description
+            description = frontmatter.get("description", "")
+            if not description:
+                description = f"Spec Kit: {command_name} workflow"
+
+            # Build SKILL.md with manually formatted frontmatter
+            def _quote(v: str) -> str:
+                escaped = v.replace("\\", "\\\\").replace('"', '\\"')
+                return f'"{escaped}"'
+
+            skill_content = (
+                f"---\n"
+                f"name: {_quote(skill_name)}\n"
+                f"description: {_quote(description)}\n"
+                f"compatibility: "
+                f"{_quote('Requires spec-kit project structure with .specify/ directory')}\n"
+                f"metadata:\n"
+                f"  author: {_quote('github-spec-kit')}\n"
+                f"  source: {_quote('templates/commands/' + src_file.name)}\n"
+                f"---\n"
+                f"{processed_body}"
+            )
+
+            skill_content = self.post_process_skill_content(skill_content)
+
+            # Write directly to global ~/.hermes/skills/speckit-<name>/SKILL.md
+            skill_dir = global_skills_dir / skill_name
+            skill_dir.mkdir(parents=True, exist_ok=True)
+            skill_file = skill_dir / "SKILL.md"
+            normalized = skill_content.replace("\r\n", "\n")
+            skill_file.write_bytes(normalized.encode("utf-8"))
+            created.append(skill_file)
+
+        # Upsert managed context section into the agent context file
+        self.upsert_context_section(project_root)
+
+        # Create project-local marker directory so extension commands
+        # (e.g. git) can detect Hermes as an active integration.
+        # Hermes itself ignores this directory — skills live globally.
+        (project_root / ".hermes" / "skills").mkdir(parents=True, exist_ok=True)
+
+        return created
+
+    # -- Uninstall ---------------------------------------------------------
+
+    def teardown(
+        self,
+        project_root: Path,
+        manifest: IntegrationManifest,
+        *,
+        force: bool = False,
+    ) -> tuple[list[Path], list[Path]]:
+        """Uninstall integration files including global Hermes skills.
+
+        Removes the managed context section from AGENTS.md, removes the
+        project-local marker directory (if empty), delegates to
+        ``manifest.uninstall()`` for project-local tracked files, and
+        removes all ``speckit-*`` skills under ``~/.hermes/skills/``.
+
+        Global skills are always removed on teardown — this matches the
+        standard integration behaviour where all files created by the
+        integration are removed on ``specify integration uninstall``.
+        """
+        # Remove managed context section from AGENTS.md
+        self.remove_context_section(project_root)
+
+        # Delegate to manifest for project-local tracked files (scripts,
+        # templates, context entries tracked in the manifest).
+        removed, skipped = manifest.uninstall(project_root, force=force)
+
+        # Remove project-local marker directory if empty
+        local_skills_dir = project_root / ".hermes" / "skills"
+        if local_skills_dir.is_dir() and not any(local_skills_dir.iterdir()):
+            local_skills_dir.rmdir()
+            hermes_dir = project_root / ".hermes"
+            if hermes_dir.is_dir() and not any(hermes_dir.iterdir()):
+                hermes_dir.rmdir()
+
+        # Remove all global Hermes skills for speckit — these are always
+        # removed on uninstall regardless of the force flag, matching the
+        # standard behaviour where all integration files are cleaned up.
+        global_skills_dir = self._hermes_home_skills_dir()
+        if global_skills_dir.is_dir():
+            for skill_dir in sorted(global_skills_dir.iterdir()):
+                if skill_dir.is_dir() and skill_dir.name.startswith("speckit-"):
+                    try:
+                        rmtree(skill_dir)
+                        removed.append(skill_dir)
+                    except OSError:
+                        skipped.append(skill_dir)
+
+        return removed, skipped
+
+    # -- CLI dispatch ------------------------------------------------------
+
+    def build_exec_args(
+        self,
+        prompt: str,
+        *,
+        model: str | None = None,
+        output_json: bool = True,
+    ) -> list[str] | None:
+        """Build Hermes CLI invocation for programmatic dispatch.
+
+        Uses ``hermes chat -Q -q`` for one-shot queries in quiet mode,
+        mapping slash-command invocations to the appropriate skill-based
+        dispatch.
+        """
+        args = [self._resolve_executable(), "chat", "-Q"]
+
+        if model:
+            args.extend(["-m", model])
+        if output_json:
+            args.append("--json")
+
+        # If prompt starts with a slash command, pass it directly
+        # so Hermes can dispatch to the appropriate skill.
+        if prompt.startswith("/"):
+            command, _, remainder = prompt[1:].partition(" ")
+            if command:
+                args.extend(["-s", command])
+                if remainder:
+                    args.extend(["-q", remainder])
+            else:
+                args.extend(["-q", prompt])
+        else:
+            args.extend(["-q", prompt])
+
+        return args

src/specify_cli/integrations/opencode/__init__.py

@@ -28,7 +28,12 @@ class OpencodeIntegration(MarkdownIntegration):
         model: str | None = None,
         output_json: bool = True,
     ) -> list[str] | None:
-        args = [self.key, "run"]
+        args = [self._resolve_executable(), "run"]
+        # Apply operator-injected extra args before the prompt-derived
+        # --command and the canonical --format/-m flags so Spec Kit's
+        # later appends remain authoritative under repeated-flag CLI
+        # semantics.
+        self._apply_extra_args_env_var(args)
 
         message = prompt
         if prompt.startswith("/"):

src/specify_cli/integrations/vibe/__init__.py

@@ -81,13 +81,13 @@ class VibeIntegration(SkillsIntegration):
             out.append(line)
         return "".join(out)
 
-
     def post_process_skill_content(self, content: str) -> str:
         """
-        Inject Vibe-specific frontmatter flags:
+        Inject shared hook guidance and Vibe-specific frontmatter flags:
         - user-invocable: allows the skill to be invoked by the user (not just other agents)
         """
-        updated = self._inject_frontmatter_flag(content, "user-invocable")
+        updated = super().post_process_skill_content(content)
+        updated = self._inject_frontmatter_flag(updated, "user-invocable")
         return updated
 
     def setup(
@@ -107,27 +107,4 @@ class VibeIntegration(SkillsIntegration):
             err=True,
         )
 
-        created = super().setup(project_root, manifest, parsed_options=parsed_options, **opts)
-
-        # Post-process generated skill files
-        skills_dir = self.skills_dest(project_root).resolve()
-
-        for path in created:
-            # Only touch SKILL.md files under the skills directory
-            try:
-                path.resolve().relative_to(skills_dir)
-            except ValueError:
-                continue
-            if path.name != "SKILL.md":
-                continue
-
-            content_bytes = path.read_bytes()
-            content = content_bytes.decode("utf-8")
-
-            updated = self.post_process_skill_content(content)
-
-            if updated != content:
-                path.write_bytes(updated.encode("utf-8"))
-                self.record_file_in_manifest(path, project_root, manifest)
-
-        return created
+        return super().setup(project_root, manifest, parsed_options=parsed_options, **opts)

src/specify_cli/presets.py

@@ -28,6 +28,7 @@ from packaging import version as pkg_version
 from packaging.specifiers import SpecifierSet, InvalidSpecifier
 
 from .extensions import REINSTALL_COMMAND, ExtensionRegistry, normalize_priority
+from .integrations.base import IntegrationBase
 
 
 def _substitute_core_template(
@@ -1058,6 +1059,9 @@ class PresetManager:
                         body = registrar.resolve_skill_placeholders(
                             selected_ai, fm, body, self.project_root
                         )
+                        body = self._resolve_skill_command_refs(
+                            body, registrar, selected_ai
+                        )
                     fm_data = registrar.build_skill_frontmatter(
                         selected_ai if isinstance(selected_ai, str) else "",
                         skill_name, desc,
@@ -1097,37 +1101,24 @@ class PresetManager:
     def _get_skills_dir(self) -> Optional[Path]:
         """Return the active skills directory for preset skill overrides.
 
-        Reads ``.specify/init-options.json`` to determine whether skills
-        are enabled and which agent was selected, then delegates to
-        the module-level ``_get_skills_dir()`` helper for the concrete path.
+        Delegates to :func:`resolve_active_skills_dir` which reads
+        init-options, applies the Kimi native-skills fallback, and
+        safely creates the directory when ``ai_skills`` is enabled.
 
-        Kimi is treated as a native-skills agent: if ``ai == "kimi"`` and
-        ``.kimi/skills`` exists, presets should still propagate command
-        overrides to skills even when ``ai_skills`` is false.
-
-        Returns:
-            The skills directory ``Path``, or ``None`` if skills were not
-            enabled and no native-skills fallback applies.
+        Returns ``None`` (instead of raising) when the directory cannot
+        be created due to symlink, containment, or permission issues so
+        that callers can fall back gracefully.
         """
-        from . import load_init_options, _get_skills_dir
-
-        opts = load_init_options(self.project_root)
-        if not isinstance(opts, dict):
-            opts = {}
-        agent = opts.get("ai")
-        if not isinstance(agent, str) or not agent:
+        from . import resolve_active_skills_dir, _print_cli_warning
+        try:
+            return resolve_active_skills_dir(self.project_root)
+        except (ValueError, OSError) as exc:
+            _print_cli_warning(
+                "resolve", "skills directory", None, exc,
+                continuing="Continuing without skill registration.",
+            )
             return None
 
-        ai_skills_enabled = bool(opts.get("ai_skills"))
-        if not ai_skills_enabled and agent != "kimi":
-            return None
-
-        skills_dir = _get_skills_dir(self.project_root, agent)
-        if not skills_dir.is_dir():
-            return None
-
-        return skills_dir
-
     @staticmethod
     def _skill_names_for_command(cmd_name: str) -> tuple[str, str]:
         """Return the modern and legacy skill directory names for a command."""
@@ -1147,6 +1138,23 @@ class PresetManager:
             title_name = title_name[len("speckit."):]
         return title_name.replace(".", " ").replace("-", " ").title()
 
+    @staticmethod
+    def _resolve_skill_command_refs(
+        body: str, registrar: "CommandRegistrar", selected_ai: str
+    ) -> str:
+        """Render ``__SPECKIT_COMMAND_*__`` tokens in a skill body as invocations.
+
+        Looks up the agent's invoke separator and rewrites each
+        ``__SPECKIT_COMMAND_<NAME>__`` placeholder into the matching
+        slash-command invocation — ``/speckit-<cmd>`` for a ``-`` separator,
+        ``/speckit.<cmd>`` for ``.`` — the same rendering the command layer
+        applies via ``CommandRegistrar.register_commands()``.
+        """
+        separator = registrar.AGENT_CONFIGS.get(selected_ai, {}).get(
+            "invoke_separator", "."
+        )
+        return IntegrationBase.resolve_command_refs(body, separator)
+
     def _build_extension_skill_restore_index(self) -> Dict[str, Dict[str, Any]]:
         """Index extension-backed skill restore data by skill directory name."""
         from .extensions import ExtensionManifest, ValidationError
@@ -1323,6 +1331,7 @@ class PresetManager:
             body = registrar.resolve_skill_placeholders(
                 selected_ai, frontmatter, body, self.project_root
             )
+            body = self._resolve_skill_command_refs(body, registrar, selected_ai)
 
             for target_skill_name in target_skill_names:
                 skill_subdir = skills_dir / target_skill_name
@@ -1415,6 +1424,9 @@ class PresetManager:
                     body = registrar.resolve_skill_placeholders(
                         selected_ai, frontmatter, body, self.project_root
                     )
+                    body = self._resolve_skill_command_refs(
+                        body, registrar, selected_ai
+                    )
 
                 original_desc = frontmatter.get("description", "")
                 enhanced_desc = original_desc or SKILL_DESCRIPTIONS.get(
@@ -1452,6 +1464,9 @@ class PresetManager:
                     body = registrar.resolve_skill_placeholders(
                         selected_ai, frontmatter, body, self.project_root
                     )
+                    body = self._resolve_skill_command_refs(
+                        body, registrar, selected_ai
+                    )
 
                 command_name = extension_restore["command_name"]
                 title_name = self._skill_title_from_command(command_name)
@@ -1903,12 +1918,24 @@ class PresetCatalog:
             if not url:
                 continue
             self._validate_catalog_url(url)
+            raw_priority = item.get("priority", idx + 1)
+            # Reject bools explicitly: ``bool`` is a subclass of ``int`` so
+            # ``int(True)`` silently returns 1, which would let a YAML
+            # ``priority: true`` slip through as a valid priority of 1. The
+            # sibling integration-catalog reader in ``catalogs.py`` already
+            # guards this; mirror the check here so the three catalog
+            # validators stay consistent.
+            if isinstance(raw_priority, bool):
+                raise PresetValidationError(
+                    f"Invalid priority for catalog '{item.get('name', idx + 1)}': "
+                    f"expected integer, got {raw_priority!r}"
+                )
             try:
-                priority = int(item.get("priority", idx + 1))
+                priority = int(raw_priority)
             except (TypeError, ValueError):
                 raise PresetValidationError(
                     f"Invalid priority for catalog '{item.get('name', idx + 1)}': "
-                    f"expected integer, got {item.get('priority')!r}"
+                    f"expected integer, got {raw_priority!r}"
                 )
             raw_install = item.get("install_allowed", False)
             if isinstance(raw_install, str):

src/specify_cli/shared_infra.py

@@ -88,7 +88,13 @@ def _shared_relative_path(project_path: Path, dest: Path) -> Path:
     return rel
 
 
-def _ensure_safe_shared_directory(project_path: Path, directory: Path, *, create: bool = True) -> None:
+def _ensure_safe_shared_directory(
+    project_path: Path,
+    directory: Path,
+    *,
+    create: bool = True,
+    context: str = "shared infrastructure directory",
+) -> None:
     """Create a shared infra directory without following symlinked parents."""
     root = project_path.resolve()
     rel = _shared_relative_path(project_path, directory)
@@ -98,24 +104,24 @@ def _ensure_safe_shared_directory(project_path: Path, directory: Path, *, create
         current = current / part
         label = _shared_destination_label(project_path, current)
         if current.is_symlink():
-            raise SymlinkedSharedPathError(f"Refusing to use symlinked shared infrastructure directory: {label}")
+            raise SymlinkedSharedPathError(f"Refusing to use symlinked {context}: {label}")
         if current.exists():
             if not current.is_dir():
-                raise ValueError(f"Shared infrastructure directory path is not a directory: {label}")
+                raise ValueError(f"{context.capitalize()} path is not a directory: {label}")
             try:
                 current.resolve().relative_to(root)
             except (OSError, ValueError):
-                raise ValueError(f"Shared infrastructure directory escapes project root: {label}") from None
+                raise ValueError(f"{context.capitalize()} escapes project root: {label}") from None
             continue
         if not create:
-            raise ValueError(f"Shared infrastructure directory does not exist: {label}")
+            raise ValueError(f"{context.capitalize()} does not exist: {label}")
         current.mkdir()
         if current.is_symlink():
-            raise SymlinkedSharedPathError(f"Refusing to use symlinked shared infrastructure directory: {label}")
+            raise SymlinkedSharedPathError(f"Refusing to use symlinked {context}: {label}")
         try:
             current.resolve().relative_to(root)
         except (OSError, ValueError):
-            raise ValueError(f"Shared infrastructure directory escapes project root: {label}") from None
+            raise ValueError(f"{context.capitalize()} escapes project root: {label}") from None
 
 
 def _validate_safe_shared_directory(project_path: Path, directory: Path) -> None:
@@ -363,7 +369,16 @@ def install_shared_infra(
 
                         if not _ensure_or_bucket_dir(dst_path.parent):
                             continue
-                        planned_copies.append((dst_path, rel, src_path.read_bytes(), src_path.stat().st_mode & 0o777))
+                        content = src_path.read_text(encoding="utf-8")
+                        content = IntegrationBase.resolve_command_refs(content, invoke_separator)
+                        planned_copies.append(
+                            (
+                                dst_path,
+                                rel,
+                                content.encode("utf-8"),
+                                src_path.stat().st_mode & 0o777,
+                            )
+                        )
 
     templates_src = shared_templates_source(core_pack=core_pack, repo_root=repo_root)
     if templates_src.is_dir():

src/specify_cli/workflows/engine.py

@@ -11,6 +11,7 @@ The engine is the orchestrator that:
 from __future__ import annotations
 
 import json
+import os
 import re
 import uuid
 from datetime import datetime, timezone
@@ -425,7 +426,7 @@ class WorkflowEngine:
         inputs:
             User-provided input values.
         run_id:
-            Optional run ID (auto-generated if not provided).
+            Optional run ID (uses SPECKIT_WORKFLOW_RUN_ID when set, otherwise auto-generated).
 
         Returns
         -------
@@ -433,8 +434,14 @@ class WorkflowEngine:
         """
         from . import STEP_REGISTRY
 
+        effective_run_id = run_id
+        if effective_run_id is None:
+            env_run_id = os.environ.get("SPECKIT_WORKFLOW_RUN_ID", "").strip()
+            if env_run_id:
+                effective_run_id = env_run_id
+
         state = RunState(
-            run_id=run_id,
+            run_id=effective_run_id,
             workflow_id=definition.id,
             project_root=self.project_root,
         )
@@ -673,22 +680,29 @@ class WorkflowEngine:
                         if not evaluate_condition(condition, context):
                             break
                         # Namespace nested step IDs per iteration
-                        iter_steps = []
-                        for ns in result.next_steps:
+                        # so logs and state keys are unique.
+                        # Execute one step at a time and alias each
+                        # result back to the unprefixed key so that
+                        # later steps in the same body and the loop
+                        # condition see the latest values.
+                        for ns_idx, ns in enumerate(result.next_steps):
                             ns_copy = dict(ns)
-                            if "id" in ns_copy:
-                                ns_copy["id"] = f"{step_id}:{ns_copy['id']}:{_loop_iter + 1}"
-                            iter_steps.append(ns_copy)
-                        self._execute_steps(
-                            iter_steps, context, state, registry,
-                            step_offset=-1,
-                        )
-                        if state.status in (
-                            RunStatus.PAUSED,
-                            RunStatus.FAILED,
-                            RunStatus.ABORTED,
-                        ):
-                            return
+                            orig = ns_copy.get("id")
+                            base_id = orig or f"step-{ns_idx}"
+                            ns_copy["id"] = f"{step_id}:{base_id}:{_loop_iter + 1}"
+                            self._execute_steps(
+                                [ns_copy], context, state, registry,
+                                step_offset=-1,
+                            )
+                            if state.status in (
+                                RunStatus.PAUSED,
+                                RunStatus.FAILED,
+                                RunStatus.ABORTED,
+                            ):
+                                return
+                            if orig and ns_copy["id"] in context.steps:
+                                context.steps[orig] = context.steps[ns_copy["id"]]
+                                state.step_results[orig] = context.steps[ns_copy["id"]]
 
             # Fan-out: execute nested step template per item with unique IDs
             if step_type == "fan-out":

src/specify_cli/workflows/expressions.py

@@ -102,6 +102,15 @@ def _build_namespace(context: Any) -> dict[str, Any]:
         ns["item"] = context.item
     if hasattr(context, "fan_in"):
         ns["fan_in"] = context.fan_in or {}
+    # Engine-managed runtime metadata. Always present (even outside a
+    # run) so templates referencing it never error: `run_id` falls back
+    # to an empty string when no run is active (dry-run, validation,
+    # ad-hoc evaluator usage). The value is the same one Spec Kit
+    # prints as `Run ID:` at the end of `workflow run` — auto-generated
+    # runs use an 8-character uuid4 hex; operator-supplied ids may be
+    # any alphanumeric string with hyphens or underscores.
+    run_id = getattr(context, "run_id", None) or ""
+    ns["context"] = {"run_id": run_id}
     return ns
 
 

templates/commands/clarify.md

@@ -197,13 +197,24 @@ Execution steps:
 
 7. Write the updated spec back to `FEATURE_SPEC`.
 
-8. Report completion (after questioning loop ends or early termination):
-   - Number of questions asked & answered.
-   - Path to updated spec.
-   - Sections touched (list names).
-   - Coverage summary table listing each taxonomy category with Status: Resolved (was Partial/Missing and addressed), Deferred (exceeds question quota or better suited for planning), Clear (already sufficient), Outstanding (still Partial/Missing but low impact).
-   - If any Outstanding or Deferred remain, recommend whether to proceed to `__SPECKIT_COMMAND_PLAN__` or run `__SPECKIT_COMMAND_CLARIFY__` again later post-plan.
-   - Suggested next command.
+8. **Re-validate Spec Quality Checklist** (if it exists):
+   - Check if `FEATURE_DIR/checklists/requirements.md` exists.
+   - If it does NOT exist, skip this step silently.
+   - If it exists:
+     1. Read the checklist file.
+     2. Identify all GitHub task-list checkbox lines — lines matching `- [ ]`, `- [x]`, or `- [X]` (case-insensitive, tolerant of leading whitespace for nested items) outside of code fences. Ignore all other content (headings, notes, non-checkbox bullets, metadata).
+     3. For each checkbox line, record its current marker state (checked or unchecked) and item text into a before-snapshot list.
+     4. Re-evaluate each checkbox item against the **updated** spec (the version just saved in step 7).
+     5. For each checkbox item, update only if the checked/unchecked state actually changes:
+        - If the item now passes and was unchecked: change `[ ]` to `[x]`.
+        - If the item now fails and was checked: change `[x]`/`[X]` to `[ ]`.
+        - If the state is unchanged: leave the marker as-is (preserve existing case to avoid cosmetic diffs).
+     6. Save the updated checklist file. **Only toggle the `[ ]`/`[x]` marker portion of checkbox lines whose state changed.** All other file content — headings, metadata, notes, line ordering, whitespace — must remain unchanged to avoid noisy diffs.
+     7. Compare the before-snapshot with the current state to compute three lists for the Completion Report:
+        - **Newly passing**: items that changed from unchecked to checked.
+        - **Regressions**: items that changed from checked to unchecked.
+        - **Still unchecked**: items that remain unchecked.
+     8. Record the before/after pass counts as checked/total checkbox items (e.g., "12/16 → 15/16 items passing").
 
 Behavior rules:
 
@@ -217,17 +228,27 @@ Behavior rules:
 
 Context for prioritization: {ARGS}
 
-## Post-Execution Checks
+## Mandatory Post-Execution Hooks
+
+**You MUST complete this section before reporting completion to the user.**
 
-**Check for extension hooks (after clarification)**:
 Check if `.specify/extensions.yml` exists in the project root.
-- If it exists, read it and look for entries under the `hooks.after_clarify` key
-- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
+- If it does not exist, or no hooks are registered under `hooks.after_clarify`, skip to the Completion Report.
+- If it exists, read it and look for entries under the `hooks.after_clarify` key.
+- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue to the Completion Report.
 - Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
 - For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
   - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
   - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
 - For each executable hook, output the following based on its `optional` flag:
+  - **Mandatory hook** (`optional: false`) — **You MUST emit `EXECUTE_COMMAND:` for each mandatory hook**:
+    ```
+    ## Extension Hooks
+
+    **Automatic Hook**: {extension}
+    Executing: `/{command}`
+    EXECUTE_COMMAND: {command}
+    ```
   - **Optional hook** (`optional: true`):
     ```
     ## Extension Hooks
@@ -239,12 +260,21 @@ Check if `.specify/extensions.yml` exists in the project root.
     Prompt: {prompt}
     To execute: `/{command}`
     ```
-  - **Mandatory hook** (`optional: false`):
-    ```
-    ## Extension Hooks
 
-    **Automatic Hook**: {extension}
-    Executing: `/{command}`
-    EXECUTE_COMMAND: {command}
-    ```
-- If no hooks are registered or `.specify/extensions.yml` does not exist, skip silently
+## Completion Report
+
+Report completion (after questioning loop ends or early termination):
+- Number of questions asked & answered.
+- Path to updated spec.
+- Sections touched (list names).
+- Spec quality checklist status (if `FEATURE_DIR/checklists/requirements.md` was re-validated): show before/after pass counts (e.g., "Spec Quality Checklist: 12/16 → 15/16 items passing") and list any items that changed state — both newly checked (unchecked → checked) and any regressions (checked → unchecked). If any items remain unchecked, list them as areas needing attention.
+- Coverage summary table listing each taxonomy category with Status: Resolved (was Partial/Missing and addressed), Deferred (exceeds question quota or better suited for planning), Clear (already sufficient), Outstanding (still Partial/Missing but low impact).
+- If any Outstanding or Deferred remain, recommend whether to proceed to `__SPECKIT_COMMAND_PLAN__` or run `__SPECKIT_COMMAND_CLARIFY__` again later post-plan.
+- Suggested next command.
+
+## Done When
+
+- [ ] Spec ambiguities identified and clarifications integrated into spec file
+- [ ] Spec quality checklist re-validated against updated spec (if `FEATURE_DIR/checklists/requirements.md` exists)
+- [ ] Extension hooks dispatched or skipped according to the rules in Mandatory Post-Execution Hooks above
+- [ ] Completion reported to user with questions answered, sections touched, checklist status, and coverage summary

templates/commands/implement.md

@@ -168,35 +168,49 @@ You **MUST** consider the user input before proceeding (if not empty).
    - Check that implemented features match the original specification
    - Validate that tests pass and coverage meets requirements
    - Confirm the implementation follows the technical plan
-   - Report final status with summary of completed work
 
 Note: This command assumes a complete task breakdown exists in tasks.md. If tasks are incomplete or missing, suggest running `__SPECKIT_COMMAND_TASKS__` first to regenerate the task list.
 
-10. **Check for extension hooks**: After completion validation, check if `.specify/extensions.yml` exists in the project root.
-    - If it exists, read it and look for entries under the `hooks.after_implement` key
-    - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
-    - Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
-    - For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
-      - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
-      - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
-    - For each executable hook, output the following based on its `optional` flag:
-      - **Optional hook** (`optional: true`):
-        ```
-        ## Extension Hooks
+## Mandatory Post-Execution Hooks
 
-        **Optional Hook**: {extension}
-        Command: `/{command}`
-        Description: {description}
+**You MUST complete this section before reporting completion to the user.**
 
-        Prompt: {prompt}
-        To execute: `/{command}`
-        ```
-      - **Mandatory hook** (`optional: false`):
-        ```
-        ## Extension Hooks
+Check if `.specify/extensions.yml` exists in the project root.
+- If it does not exist, or no hooks are registered under `hooks.after_implement`, skip to the Completion Report.
+- If it exists, read it and look for entries under the `hooks.after_implement` key.
+- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue to the Completion Report.
+- Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
+- For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
+  - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
+  - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
+- For each executable hook, output the following based on its `optional` flag:
+  - **Mandatory hook** (`optional: false`) — **You MUST emit `EXECUTE_COMMAND:` for each mandatory hook**:
+    ```
+    ## Extension Hooks
 
-        **Automatic Hook**: {extension}
-        Executing: `/{command}`
-        EXECUTE_COMMAND: {command}
-        ```
-    - If no hooks are registered or `.specify/extensions.yml` does not exist, skip silently
+    **Automatic Hook**: {extension}
+    Executing: `/{command}`
+    EXECUTE_COMMAND: {command}
+    ```
+  - **Optional hook** (`optional: true`):
+    ```
+    ## Extension Hooks
+
+    **Optional Hook**: {extension}
+    Command: `/{command}`
+    Description: {description}
+
+    Prompt: {prompt}
+    To execute: `/{command}`
+    ```
+
+## Completion Report
+
+Report final status with summary of completed work.
+
+## Done When
+
+- [ ] All tasks in tasks.md completed and marked `[X]`
+- [ ] Implementation validated against specification, plan, and test coverage
+- [ ] Extension hooks dispatched or skipped according to the rules in Mandatory Post-Execution Hooks above
+- [ ] Completion reported to user with summary of completed work

templates/commands/plan.md

@@ -70,36 +70,42 @@ You **MUST** consider the user input before proceeding (if not empty).
    - Phase 1: Update agent context by running the agent script
    - Re-evaluate Constitution Check post-design
 
-4. **Stop and report**: Command ends after Phase 2 planning. Report branch, IMPL_PLAN path, and generated artifacts.
+## Mandatory Post-Execution Hooks
 
-5. **Check for extension hooks**: After reporting, check if `.specify/extensions.yml` exists in the project root.
-   - If it exists, read it and look for entries under the `hooks.after_plan` key
-   - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
-   - Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
-   - For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
-     - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
-     - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
-   - For each executable hook, output the following based on its `optional` flag:
-     - **Optional hook** (`optional: true`):
-       ```
-       ## Extension Hooks
+**You MUST complete this section before reporting completion to the user.**
 
-       **Optional Hook**: {extension}
-       Command: `/{command}`
-       Description: {description}
+Check if `.specify/extensions.yml` exists in the project root.
+- If it does not exist, or no hooks are registered under `hooks.after_plan`, skip to the Completion Report.
+- If it exists, read it and look for entries under the `hooks.after_plan` key.
+- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue to the Completion Report.
+- Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
+- For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
+  - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
+  - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
+- For each executable hook, output the following based on its `optional` flag:
+  - **Mandatory hook** (`optional: false`) — **You MUST emit `EXECUTE_COMMAND:` for each mandatory hook**:
+    ```
+    ## Extension Hooks
 
-       Prompt: {prompt}
-       To execute: `/{command}`
-       ```
-     - **Mandatory hook** (`optional: false`):
-       ```
-       ## Extension Hooks
+    **Automatic Hook**: {extension}
+    Executing: `/{command}`
+    EXECUTE_COMMAND: {command}
+    ```
+  - **Optional hook** (`optional: true`):
+    ```
+    ## Extension Hooks
 
-       **Automatic Hook**: {extension}
-       Executing: `/{command}`
-       EXECUTE_COMMAND: {command}
-       ```
-   - If no hooks are registered or `.specify/extensions.yml` does not exist, skip silently
+    **Optional Hook**: {extension}
+    Command: `/{command}`
+    Description: {description}
+
+    Prompt: {prompt}
+    To execute: `/{command}`
+    ```
+
+## Completion Report
+
+Command ends after Phase 2 planning. Report branch, IMPL_PLAN path, and generated artifacts.
 
 ## Phases
 
@@ -150,3 +156,9 @@ You **MUST** consider the user input before proceeding (if not empty).
 
 - Use absolute paths for filesystem operations; use project-relative paths for references in documentation and agent context files
 - ERROR on gate failures or unresolved clarifications
+
+## Done When
+
+- [ ] Plan workflow executed and design artifacts generated
+- [ ] Extension hooks dispatched or skipped according to the rules in Mandatory Post-Execution Hooks above
+- [ ] Completion reported to user with branch, plan path, and generated artifacts

templates/commands/specify.md

@@ -183,7 +183,7 @@ Given that feature description, do this:
 
    c. **Handle Validation Results**:
 
-      - **If all items pass**: Mark checklist complete and proceed to step 8
+      - **If all items pass**: Mark checklist complete and proceed to the Mandatory Post-Execution Hooks section
 
       - **If items fail (excluding [NEEDS CLARIFICATION])**:
         1. List the failing items and specific issues
@@ -228,40 +228,46 @@ Given that feature description, do this:
 
    d. **Update Checklist**: After each validation iteration, update the checklist file with current pass/fail status
 
-8. **Report completion** to the user with:
-   - `SPECIFY_FEATURE_DIRECTORY` — the feature directory path
-   - `SPEC_FILE` — the spec file path
-   - Checklist results summary
-   - Readiness for the next phase (`__SPECKIT_COMMAND_CLARIFY__` or `__SPECKIT_COMMAND_PLAN__`)
+## Mandatory Post-Execution Hooks
 
-9. **Check for extension hooks**: After reporting completion, check if `.specify/extensions.yml` exists in the project root.
-   - If it exists, read it and look for entries under the `hooks.after_specify` key
-   - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
-   - Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
-   - For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
-     - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
-     - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
-   - For each executable hook, output the following based on its `optional` flag:
-     - **Optional hook** (`optional: true`):
-       ```
-       ## Extension Hooks
+**You MUST complete this section before reporting completion to the user.**
 
-       **Optional Hook**: {extension}
-       Command: `/{command}`
-       Description: {description}
+Check if `.specify/extensions.yml` exists in the project root.
+- If it does not exist, or no hooks are registered under `hooks.after_specify`, skip to the Completion Report.
+- If it exists, read it and look for entries under the `hooks.after_specify` key.
+- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue to the Completion Report.
+- Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
+- For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
+  - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
+  - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
+- For each executable hook, output the following based on its `optional` flag:
+  - **Mandatory hook** (`optional: false`) — **You MUST emit `EXECUTE_COMMAND:` for each mandatory hook**:
+    ```
+    ## Extension Hooks
 
-       Prompt: {prompt}
-       To execute: `/{command}`
-       ```
-     - **Mandatory hook** (`optional: false`):
-       ```
-       ## Extension Hooks
+    **Automatic Hook**: {extension}
+    Executing: `/{command}`
+    EXECUTE_COMMAND: {command}
+    ```
+  - **Optional hook** (`optional: true`):
+    ```
+    ## Extension Hooks
 
-       **Automatic Hook**: {extension}
-       Executing: `/{command}`
-       EXECUTE_COMMAND: {command}
-       ```
-   - If no hooks are registered or `.specify/extensions.yml` does not exist, skip silently
+    **Optional Hook**: {extension}
+    Command: `/{command}`
+    Description: {description}
+
+    Prompt: {prompt}
+    To execute: `/{command}`
+    ```
+
+## Completion Report
+
+Report completion to the user with:
+- `SPECIFY_FEATURE_DIRECTORY` — the feature directory path
+- `SPEC_FILE` — the spec file path
+- Checklist results summary
+- Readiness for the next phase (`__SPECKIT_COMMAND_CLARIFY__` or `__SPECKIT_COMMAND_PLAN__`)
 
 **NOTE:** Branch creation is handled by the `before_specify` hook (git extension). Spec directory and file creation are always handled by this core command.
 
@@ -325,3 +331,9 @@ Success criteria must be:
 - "Database can handle 1000 TPS" (implementation detail, use user-facing metric)
 - "React components render efficiently" (framework-specific)
 - "Redis cache hit rate above 80%" (technology-specific)
+
+## Done When
+
+- [ ] Specification written to `SPEC_FILE` and validated against quality checklist
+- [ ] Extension hooks dispatched or skipped according to the rules in Mandatory Post-Execution Hooks above
+- [ ] Completion reported to user with feature directory, spec file path, and checklist results

templates/commands/tasks.md

@@ -89,42 +89,48 @@ You **MUST** consider the user input before proceeding (if not empty).
    - Parallel execution examples per story
    - Implementation strategy section (MVP first, incremental delivery)
 
-5. **Report**: Output path to generated tasks.md and summary:
-   - Total task count
-   - Task count per user story
-   - Parallel opportunities identified
-   - Independent test criteria for each story
-   - Suggested MVP scope (typically just User Story 1)
-   - Format validation: Confirm ALL tasks follow the checklist format (checkbox, ID, labels, file paths)
+## Mandatory Post-Execution Hooks
 
-6. **Check for extension hooks**: After tasks.md is generated, check if `.specify/extensions.yml` exists in the project root.
-   - If it exists, read it and look for entries under the `hooks.after_tasks` key
-   - If the YAML cannot be parsed or is invalid, skip hook checking silently and continue normally
-   - Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
-   - For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
-     - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
-     - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
-   - For each executable hook, output the following based on its `optional` flag:
-     - **Optional hook** (`optional: true`):
-       ```
-       ## Extension Hooks
+**You MUST complete this section before reporting completion to the user.**
 
-       **Optional Hook**: {extension}
-       Command: `/{command}`
-       Description: {description}
+Check if `.specify/extensions.yml` exists in the project root.
+- If it does not exist, or no hooks are registered under `hooks.after_tasks`, skip to the Completion Report.
+- If it exists, read it and look for entries under the `hooks.after_tasks` key.
+- If the YAML cannot be parsed or is invalid, skip hook checking silently and continue to the Completion Report.
+- Filter out hooks where `enabled` is explicitly `false`. Treat hooks without an `enabled` field as enabled by default.
+- For each remaining hook, do **not** attempt to interpret or evaluate hook `condition` expressions:
+  - If the hook has no `condition` field, or it is null/empty, treat the hook as executable
+  - If the hook defines a non-empty `condition`, skip the hook and leave condition evaluation to the HookExecutor implementation
+- For each executable hook, output the following based on its `optional` flag:
+  - **Mandatory hook** (`optional: false`) — **You MUST emit `EXECUTE_COMMAND:` for each mandatory hook**:
+    ```
+    ## Extension Hooks
 
-       Prompt: {prompt}
-       To execute: `/{command}`
-       ```
-     - **Mandatory hook** (`optional: false`):
-       ```
-       ## Extension Hooks
+    **Automatic Hook**: {extension}
+    Executing: `/{command}`
+    EXECUTE_COMMAND: {command}
+    ```
+  - **Optional hook** (`optional: true`):
+    ```
+    ## Extension Hooks
 
-       **Automatic Hook**: {extension}
-       Executing: `/{command}`
-       EXECUTE_COMMAND: {command}
-       ```
-   - If no hooks are registered or `.specify/extensions.yml` does not exist, skip silently
+    **Optional Hook**: {extension}
+    Command: `/{command}`
+    Description: {description}
+
+    Prompt: {prompt}
+    To execute: `/{command}`
+    ```
+
+## Completion Report
+
+Output path to generated tasks.md and summary:
+- Total task count
+- Task count per user story
+- Parallel opportunities identified
+- Independent test criteria for each story
+- Suggested MVP scope (typically just User Story 1)
+- Format validation: Confirm ALL tasks follow the checklist format (checkbox, ID, labels, file paths)
 
 Context for task generation: {ARGS}
 
@@ -201,3 +207,9 @@ Every task MUST strictly follow this format:
   - Within each story: Tests (if requested) → Models → Services → Endpoints → Integration
   - Each phase should be a complete, independently testable increment
 - **Final Phase**: Polish & Cross-Cutting Concerns
+
+## Done When
+
+- [ ] tasks.md generated with all phases, task IDs, and file paths
+- [ ] Extension hooks dispatched or skipped according to the rules in Mandatory Post-Execution Hooks above
+- [ ] Completion reported to user with task count, story breakdown, and MVP scope

tests/integrations/test_cli.py

@@ -22,6 +22,26 @@ def _normalize_cli_output(output: str) -> str:
     return output.strip()
 
 
+class TestCliDiagnosticFormatting:
+    def test_cli_error_detail_flattens_newlines(self):
+        import specify_cli
+
+        assert specify_cli._cli_error_detail(RuntimeError("line one\nline two")) == "line one line two"
+
+    def test_cli_error_detail_handles_empty_message(self):
+        import specify_cli
+
+        assert specify_cli._cli_error_detail(RuntimeError()) == "RuntimeError"
+
+    def test_cli_phase_label_includes_target(self):
+        import specify_cli
+
+        assert (
+            specify_cli._cli_phase_label("rollback", "integration", "codex")
+            == "rollback integration 'codex'"
+        )
+
+
 class TestInitIntegrationFlag:
     def test_integration_and_ai_mutually_exclusive(self, tmp_path):
         from typer.testing import CliRunner
@@ -174,6 +194,42 @@ class TestInitIntegrationFlag:
         assert normalized_output.index("Deprecation Warning") < normalized_output.index("Next Steps")
         assert (project / ".myagent" / "commands" / "speckit.plan.md").exists()
 
+    def test_init_optional_preset_failure_reports_target_and_continues(
+        self, tmp_path, monkeypatch
+    ):
+        from typer.testing import CliRunner
+        from specify_cli import app
+        from specify_cli.presets import PresetManager
+
+        def fail_install(self, path, version):
+            raise OSError("preset install exploded\nwith context")
+
+        monkeypatch.setattr(PresetManager, "install_from_directory", fail_install)
+
+        project = tmp_path / "init-preset-warning"
+        result = CliRunner().invoke(
+            app,
+            [
+                "init",
+                str(project),
+                "--integration",
+                "copilot",
+                "--script",
+                "sh",
+                "--no-git",
+                "--preset",
+                "lean",
+            ],
+            catch_exceptions=False,
+        )
+        normalized = _normalize_cli_output(result.output)
+
+        assert result.exit_code == 0, result.output
+        assert "Failed to install preset 'lean'" in normalized
+        assert "preset install exploded with context" in normalized
+        assert "Continuing without the optional preset" in normalized
+        assert "Project ready" in normalized
+
     def test_ai_claude_here_preserves_preexisting_commands(self, tmp_path):
         from typer.testing import CliRunner
         from specify_cli import app
@@ -279,10 +335,11 @@ class TestInitIntegrationFlag:
         _install_shared_infra(project, "sh", force=False)
 
         captured = capsys.readouterr()
-        assert "already exist and were not updated" in captured.out
-        assert "specify init --here --force" in captured.out
+        plain = strip_ansi(captured.out)
+        assert "already exist and were not updated" in plain
+        assert "specify init --here --force" in plain
         # Rich may wrap long lines; normalize whitespace for the second command
-        normalized = " ".join(captured.out.split())
+        normalized = " ".join(plain.split())
         assert "specify integration upgrade --force" in normalized
 
     def test_shared_infra_warns_when_manifest_cannot_be_loaded(self, tmp_path, capsys):
@@ -866,7 +923,23 @@ class TestGitExtensionAutoInstall:
 
 
 class TestSharedInfraCommandRefs:
-    """Verify _install_shared_infra resolves __SPECKIT_COMMAND_*__ in page templates."""
+    """Verify _install_shared_infra resolves __SPECKIT_COMMAND_*__ in shared infra."""
+
+    @staticmethod
+    def _combined_script_content(project, script_type):
+        script_dir = "bash" if script_type == "sh" else "powershell"
+        suffix = "sh" if script_type == "sh" else "ps1"
+        names = [
+            f"check-prerequisites.{suffix}",
+            f"common.{suffix}",
+            f"setup-tasks.{suffix}",
+        ]
+        return "\n".join(
+            (project / ".specify" / "scripts" / script_dir / name).read_text(
+                encoding="utf-8"
+            )
+            for name in names
+        )
 
     def test_dot_separator_in_page_templates(self, tmp_path):
         """Markdown agents get /speckit.<name> in page templates."""
@@ -911,6 +984,46 @@ class TestSharedInfraCommandRefs:
         assert "__SPECKIT_COMMAND_" not in content
         assert "/speckit-tasks" in content
 
+    @pytest.mark.parametrize("script_type", ["sh", "ps"])
+    def test_dot_separator_in_shared_scripts(self, tmp_path, script_type):
+        """Markdown agents get /speckit.<name> in shared script hints."""
+        from specify_cli import _install_shared_infra
+
+        project = tmp_path / f"dot-script-{script_type}"
+        project.mkdir()
+        (project / ".specify").mkdir()
+
+        _install_shared_infra(project, script_type, invoke_separator=".")
+
+        content = self._combined_script_content(project, script_type)
+        assert "__SPECKIT_COMMAND_" not in content
+        assert "/speckit.specify" in content
+        assert "/speckit.plan" in content
+        assert "/speckit.tasks" in content
+        assert "/speckit-specify" not in content
+        assert "/speckit-plan" not in content
+        assert "/speckit-tasks" not in content
+
+    @pytest.mark.parametrize("script_type", ["sh", "ps"])
+    def test_hyphen_separator_in_shared_scripts(self, tmp_path, script_type):
+        """Skills agents get /speckit-<name> in shared script hints."""
+        from specify_cli import _install_shared_infra
+
+        project = tmp_path / f"hyphen-script-{script_type}"
+        project.mkdir()
+        (project / ".specify").mkdir()
+
+        _install_shared_infra(project, script_type, invoke_separator="-")
+
+        content = self._combined_script_content(project, script_type)
+        assert "__SPECKIT_COMMAND_" not in content
+        assert "/speckit-specify" in content
+        assert "/speckit-plan" in content
+        assert "/speckit-tasks" in content
+        assert "/speckit.specify" not in content
+        assert "/speckit.plan" not in content
+        assert "/speckit.tasks" not in content
+
     def test_full_init_claude_resolves_page_templates(self, tmp_path):
         """Full CLI init with Claude (skills agent) produces hyphen refs in page templates."""
         from typer.testing import CliRunner
@@ -938,6 +1051,10 @@ class TestSharedInfraCommandRefs:
         assert "/speckit-plan" in content, "Claude (skills) should use /speckit-plan"
         assert "__SPECKIT_COMMAND_" not in content
 
+        script_content = self._combined_script_content(project, "sh")
+        assert "/speckit-specify" in script_content
+        assert "/speckit.specify" not in script_content
+
     def test_full_init_copilot_resolves_page_templates(self, tmp_path):
         """Full CLI init with Copilot (markdown agent) produces dot refs in page templates."""
         from typer.testing import CliRunner
@@ -965,6 +1082,10 @@ class TestSharedInfraCommandRefs:
         assert "/speckit.plan" in content, "Copilot (markdown) should use /speckit.plan"
         assert "__SPECKIT_COMMAND_" not in content
 
+        script_content = self._combined_script_content(project, "sh")
+        assert "/speckit.specify" in script_content
+        assert "/speckit-specify" not in script_content
+
     def test_full_init_copilot_skills_resolves_page_templates(self, tmp_path):
         """Full CLI init with Copilot --skills produces hyphen refs in page templates."""
         from typer.testing import CliRunner
@@ -994,6 +1115,10 @@ class TestSharedInfraCommandRefs:
         assert "/speckit.plan" not in content, "dot-notation leaked into Copilot skills page template"
         assert "__SPECKIT_COMMAND_" not in content
 
+        script_content = self._combined_script_content(project, "sh")
+        assert "/speckit-specify" in script_content
+        assert "/speckit.specify" not in script_content
+
 
 class TestIntegrationCatalogDiscoveryCLI:
     """End-to-end CLI tests for `integration search`, `info`, and `catalog …`.
@@ -1055,6 +1180,143 @@ class TestIntegrationCatalogDiscoveryCLI:
         finally:
             os.chdir(old)
 
+    def test_integration_install_failure_reports_phase_target_and_rollback(
+        self, tmp_path, monkeypatch
+    ):
+        from specify_cli.integrations import INTEGRATION_REGISTRY
+        from specify_cli.integrations.base import IntegrationBase
+
+        class BrokenIntegration(IntegrationBase):
+            key = "broken-test"
+            config = {
+                "name": "Broken Test",
+                "folder": ".broken/",
+                "commands_subdir": "commands",
+                "install_url": None,
+                "requires_cli": False,
+            }
+            registrar_config = {
+                "dir": ".broken/commands",
+                "format": "markdown",
+                "args": "$ARGUMENTS",
+                "extension": ".md",
+            }
+            context_file = "BROKEN.md"
+
+            def setup(self, project_root, manifest, **kwargs):
+                raise OSError("setup exploded\nwith context")
+
+            def teardown(self, project_root, manifest, force=False):
+                raise OSError("rollback exploded")
+
+        project = self._make_project(tmp_path)
+        monkeypatch.setitem(INTEGRATION_REGISTRY, "broken-test", BrokenIntegration())
+
+        result = self._invoke(["integration", "install", "broken-test"], project)
+        normalized = _normalize_cli_output(result.output)
+
+        assert result.exit_code == 1, result.output
+        assert "Failed to rollback integration 'broken-test'" in normalized
+        assert "rollback exploded" in normalized
+        assert "Failed to install integration 'broken-test'" in normalized
+        assert "setup exploded with context" in normalized
+
+    def test_integration_upgrade_failure_reports_phase_and_target(
+        self, tmp_path, monkeypatch
+    ):
+        from specify_cli.integrations import INTEGRATION_REGISTRY
+        from specify_cli.integrations.copilot import CopilotIntegration
+
+        class UpgradeBrokenIntegration(CopilotIntegration):
+            key = "upgrade-broken"
+            config = dict(CopilotIntegration.config)
+            config["name"] = "Upgrade Broken"
+
+            def setup(self, project_root, manifest, **kwargs):
+                raise OSError("upgrade exploded\nwith context")
+
+        project = self._make_project(tmp_path)
+        monkeypatch.setitem(
+            INTEGRATION_REGISTRY, "upgrade-broken", UpgradeBrokenIntegration()
+        )
+
+        (project / ".specify" / "integrations").mkdir(parents=True, exist_ok=True)
+        (project / ".specify" / "integration.json").write_text(
+            json.dumps(
+                {
+                    "version": 1,
+                    "integration": "upgrade-broken",
+                    "integrations": ["upgrade-broken"],
+                    "integration_settings": {"upgrade-broken": {"script": "sh"}},
+                }
+            ),
+            encoding="utf-8",
+        )
+        (
+            project / ".specify" / "integrations" / "upgrade-broken.manifest.json"
+        ).write_text(
+            json.dumps(
+                {
+                    "integration": "upgrade-broken",
+                    "version": "0.0.0",
+                    "installed_at": "2026-05-16T00:00:00+00:00",
+                    "files": {},
+                }
+            ),
+            encoding="utf-8",
+        )
+
+        result = self._invoke(["integration", "upgrade", "upgrade-broken"], project)
+        normalized = _normalize_cli_output(result.output)
+
+        assert result.exit_code == 1, result.output
+        assert "Failed to upgrade integration 'upgrade-broken'" in normalized
+        assert "upgrade exploded with context" in normalized
+        assert "previous integration files may still be in place" in normalized
+
+    def test_integration_switch_cleanup_warning_reports_phase_and_targets(
+        self, tmp_path, monkeypatch
+    ):
+        from specify_cli.extensions import ExtensionManager
+
+        project = self._make_project(tmp_path)
+        (project / ".specify" / "integrations").mkdir(parents=True, exist_ok=True)
+        (project / ".specify" / "integration.json").write_text(
+            json.dumps(
+                {
+                    "version": 1,
+                    "integration": "copilot",
+                    "integrations": ["copilot"],
+                    "integration_settings": {"copilot": {"script": "sh"}},
+                }
+            ),
+            encoding="utf-8",
+        )
+        (project / ".specify" / "integrations" / "copilot.manifest.json").write_text(
+            json.dumps(
+                {
+                    "integration": "copilot",
+                    "version": "0.0.0",
+                    "installed_at": "2026-05-16T00:00:00+00:00",
+                    "files": {},
+                }
+            ),
+            encoding="utf-8",
+        )
+
+        def fail_cleanup(self, integration_key):
+            raise OSError("cleanup exploded")
+
+        monkeypatch.setattr(ExtensionManager, "unregister_agent_artifacts", fail_cleanup)
+
+        result = self._invoke(["integration", "switch", "claude"], project)
+        normalized = _normalize_cli_output(result.output)
+
+        assert result.exit_code == 0, result.output
+        assert "Failed to clean up extension artifacts for integration 'copilot'" in normalized
+        assert "cleanup exploded" in normalized
+        assert "Switched to integration" in normalized
+
     # -- Project guard -----------------------------------------------------
 
     def test_search_requires_specify_project(self, tmp_path):

tests/integrations/test_extra_args.py

@@ -0,0 +1,638 @@
+"""Tests for the per-integration `SPECKIT_INTEGRATION_<KEY>_EXTRA_ARGS` and
+`SPECKIT_INTEGRATION_<KEY>_EXECUTABLE` env-var hooks.
+
+The hooks are implemented in `IntegrationBase._apply_extra_args_env_var` and
+`IntegrationBase._resolve_executable` and wired into every concrete
+`build_exec_args` — `MarkdownIntegration`, `TomlIntegration`,
+`SkillsIntegration`, plus override integrations.
+These tests cover both the shared mechanisms (via `SkillsIntegration` stubs
+near the top of the file) and override integrations end-to-end (further down).
+See issues #2595 and #2596."""
+
+import os
+
+import pytest
+
+from specify_cli.integrations.base import (
+    MarkdownIntegration,
+    SkillsIntegration,
+    TomlIntegration,
+)
+
+
+class _ClaudeStub(SkillsIntegration):
+    """Minimal Claude-like SkillsIntegration for testing."""
+
+    key = "claude"
+    config = {
+        "name": "Claude (test stub)",
+        "folder": ".claude/",
+        "commands_subdir": "skills",
+        "install_url": None,
+        "requires_cli": True,
+    }
+    registrar_config = {
+        "dir": ".claude/skills",
+        "format": "markdown",
+        "args": "$ARGUMENTS",
+        "extension": "/SKILL.md",
+    }
+    context_file = "CLAUDE.md"
+
+
+class _KiroCliStub(SkillsIntegration):
+    """SkillsIntegration with a hyphenated key to exercise key
+    normalization (`kiro-cli` → `KIRO_CLI`)."""
+
+    key = "kiro-cli"
+    config = {
+        "name": "Kiro CLI (test stub)",
+        "folder": ".kiro/",
+        "commands_subdir": "commands",
+        "install_url": None,
+        "requires_cli": True,
+    }
+    registrar_config = {
+        "dir": ".kiro/commands",
+        "format": "markdown",
+        "args": "$ARGUMENTS",
+        "extension": ".md",
+    }
+    context_file = "KIRO.md"
+
+
+class _NoCliStub(SkillsIntegration):
+    """SkillsIntegration with requires_cli=False — build_exec_args
+    must return None and the env-var hook must not fire."""
+
+    key = "no-cli"
+    config = {
+        "name": "No-CLI agent (test stub)",
+        "folder": ".no-cli/",
+        "commands_subdir": "commands",
+        "install_url": None,
+        "requires_cli": False,
+    }
+    registrar_config = {
+        "dir": ".no-cli/commands",
+        "format": "markdown",
+        "args": "$ARGUMENTS",
+        "extension": ".md",
+    }
+    context_file = "NOCLI.md"
+
+
+class _MarkdownAgentStub(MarkdownIntegration):
+    """Bare MarkdownIntegration subclass — does NOT override
+    `build_exec_args`. Locks the base implementation in
+    `MarkdownIntegration.build_exec_args` for the common case
+    (most concrete integrations: Amp, Auggie, Generic, …)."""
+
+    key = "md-agent"
+    config = {
+        "name": "Markdown agent (test stub)",
+        "folder": ".md-agent/",
+        "commands_subdir": "commands",
+        "install_url": None,
+        "requires_cli": True,
+    }
+    registrar_config = {
+        "dir": ".md-agent/commands",
+        "format": "markdown",
+        "args": "$ARGUMENTS",
+        "extension": ".md",
+    }
+    context_file = "MDAGENT.md"
+
+
+class _TomlAgentStub(TomlIntegration):
+    """Bare TomlIntegration subclass — does NOT override
+    `build_exec_args`. Locks the base implementation in
+    `TomlIntegration.build_exec_args` (Gemini, Tabnine)."""
+
+    key = "toml-agent"
+    config = {
+        "name": "TOML agent (test stub)",
+        "folder": ".toml-agent/",
+        "commands_subdir": "commands",
+        "install_url": None,
+        "requires_cli": True,
+    }
+    registrar_config = {
+        "dir": ".toml-agent/commands",
+        "format": "toml",
+        "args": "$ARGUMENTS",
+        "extension": ".toml",
+    }
+    context_file = "TOMLAGENT.md"
+
+
+@pytest.fixture(autouse=True)
+def _clean_extra_args_env(monkeypatch):
+    """Strip any leaked SPECKIT_INTEGRATION_*_EXTRA_ARGS and
+    SPECKIT_INTEGRATION_*_EXECUTABLE vars from the test env so a
+    developer's shell setting doesn't pollute results."""
+    for key in list(os.environ):
+        if key.startswith("SPECKIT_INTEGRATION_") and (
+            key.endswith("_EXTRA_ARGS") or key.endswith("_EXECUTABLE")
+        ):
+            monkeypatch.delenv(key, raising=False)
+
+
+def test_env_var_unset_byte_identical_argv():
+    """Default behaviour: env var unset → no extra args inserted.
+
+    Locks the backward-compatibility guarantee that existing
+    operators see no change.
+    """
+    args = _ClaudeStub().build_exec_args("hello prompt")
+    assert args == ["claude", "-p", "hello prompt", "--output-format", "json"]
+
+
+def test_env_var_set_flag_inserted_before_model_and_output_format(
+    monkeypatch,
+):
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS", "--dangerously-skip-permissions"
+    )
+    args = _ClaudeStub().build_exec_args("hello prompt", model="sonnet")
+    assert args == [
+        "claude",
+        "-p",
+        "hello prompt",
+        "--dangerously-skip-permissions",
+        "--model",
+        "sonnet",
+        "--output-format",
+        "json",
+    ]
+
+
+def test_env_var_multi_token_parsed_via_shlex(monkeypatch):
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS",
+        "--dangerously-skip-permissions --max-turns 3",
+    )
+    args = _ClaudeStub().build_exec_args("p")
+    assert args == [
+        "claude",
+        "-p",
+        "p",
+        "--dangerously-skip-permissions",
+        "--max-turns",
+        "3",
+        "--output-format",
+        "json",
+    ]
+
+
+def test_malformed_quoting_raises_actionable_value_error(monkeypatch):
+    """An unmatched quote in the env-var value must surface a clear
+    error naming the offending env var and showing the invalid value,
+    rather than crashing workflow dispatch with a bare shlex traceback."""
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS",
+        '--flag "unterminated',
+    )
+    with pytest.raises(ValueError) as excinfo:
+        _ClaudeStub().build_exec_args("p")
+    msg = str(excinfo.value)
+    assert "SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS" in msg
+    assert "--flag \"unterminated" in msg
+
+
+def test_env_var_empty_or_whitespace_is_noop(monkeypatch):
+    """An env var set to '' or '   ' is treated as unset."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS", "   ")
+    args = _ClaudeStub().build_exec_args("p")
+    assert args == ["claude", "-p", "p", "--output-format", "json"]
+
+
+def test_other_integration_env_var_ignored(monkeypatch):
+    """`SPECKIT_INTEGRATION_GEMINI_EXTRA_ARGS` set must NOT leak into
+    Claude's argv (per-integration scoping)."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_GEMINI_EXTRA_ARGS", "--gemini-only-flag")
+    args = _ClaudeStub().build_exec_args("p")
+    assert args == ["claude", "-p", "p", "--output-format", "json"]
+
+
+def test_key_normalization_hyphen_to_underscore_uppercase(monkeypatch):
+    """`kiro-cli` key looks up `SPECKIT_INTEGRATION_KIRO_CLI_EXTRA_ARGS`
+    (hyphens replaced with underscores, then uppercased)."""
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_KIRO_CLI_EXTRA_ARGS", "--some-kiro-flag"
+    )
+    args = _KiroCliStub().build_exec_args("p")
+    assert args == [
+        "kiro-cli",
+        "-p",
+        "p",
+        "--some-kiro-flag",
+        "--output-format",
+        "json",
+    ]
+
+
+def test_requires_cli_false_returns_none(monkeypatch):
+    """`requires_cli: False` short-circuits to None — the env-var
+    hook is never reached and no argv is built."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_NO_CLI_EXTRA_ARGS", "--should-not-appear")
+    assert _NoCliStub().build_exec_args("p") is None
+
+
+# ---------------------------------------------------------------------------
+# Base-class coverage
+#
+# Most integrations inherit `build_exec_args` from `MarkdownIntegration`
+# or `TomlIntegration` without overriding it. The tests above use
+# `SkillsIntegration` stubs (which share the same hook mechanism) — these
+# tests exercise the two other base implementations directly so all three
+# concrete bases are covered.
+# ---------------------------------------------------------------------------
+
+
+def test_markdown_integration_base_honours_extra_args(monkeypatch):
+    """A bare `MarkdownIntegration` subclass — which does not override
+    `build_exec_args` — must honour the env var via the base
+    implementation. Covers the most common integration pattern."""
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_MD_AGENT_EXTRA_ARGS", "--debug --max-tokens 100"
+    )
+    args = _MarkdownAgentStub().build_exec_args("p")
+    assert args == [
+        "md-agent",
+        "-p",
+        "p",
+        "--debug",
+        "--max-tokens",
+        "100",
+        "--output-format",
+        "json",
+    ]
+
+
+def test_toml_integration_base_honours_extra_args(monkeypatch):
+    """A bare `TomlIntegration` subclass — which does not override
+    `build_exec_args` — must honour the env var via the base
+    implementation. Covers Gemini/Tabnine-style integrations."""
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_TOML_AGENT_EXTRA_ARGS", "--yolo"
+    )
+    args = _TomlAgentStub().build_exec_args("p", model="gemini-pro")
+    # TomlIntegration uses `-m` for model (vs Markdown's `--model`).
+    assert args == [
+        "toml-agent",
+        "-p",
+        "p",
+        "--yolo",
+        "-m",
+        "gemini-pro",
+        "--output-format",
+        "json",
+    ]
+
+
+# ---------------------------------------------------------------------------
+# Override-integration coverage
+#
+# CodexIntegration, DevinIntegration, OpencodeIntegration and
+# CopilotIntegration each override `build_exec_args` rather than using the
+# base implementations. The env-var hook must be wired into every override
+# so the documented behaviour ("works for every requires_cli integration")
+# is honoured. These tests lock that contract per integration.
+# ---------------------------------------------------------------------------
+
+
+def test_codex_integration_honours_extra_args(monkeypatch):
+    from specify_cli.integrations.codex import CodexIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CODEX_EXTRA_ARGS", "--sandbox read-only")
+    args = CodexIntegration().build_exec_args("p", model="gpt-5")
+    assert args == [
+        "codex",
+        "exec",
+        "p",
+        "--sandbox",
+        "read-only",
+        "--model",
+        "gpt-5",
+        "--json",
+    ]
+
+
+def test_devin_integration_honours_extra_args(monkeypatch):
+    from specify_cli.integrations.devin import DevinIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_DEVIN_EXTRA_ARGS", "--no-confirm")
+    args = DevinIntegration().build_exec_args("p")
+    assert args == ["devin", "-p", "p", "--no-confirm"]
+
+
+def test_opencode_integration_honours_extra_args(monkeypatch):
+    from specify_cli.integrations.opencode import OpencodeIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_OPENCODE_EXTRA_ARGS", "--quiet")
+    args = OpencodeIntegration().build_exec_args("p")
+    assert args == [
+        "opencode",
+        "run",
+        "--quiet",
+        "--format",
+        "json",
+        "p",
+    ]
+
+
+def test_opencode_extra_args_cannot_clobber_prompt_derived_command(
+    monkeypatch,
+):
+    """Operator-injected extra args must appear BEFORE the prompt-derived
+    ``--command <X>`` so that Spec Kit's command selection wins under
+    repeated-flag CLI semantics (last value typically takes precedence).
+
+    Locks against the regression where an operator setting
+    ``SPECKIT_INTEGRATION_OPENCODE_EXTRA_ARGS="--command malicious"`` could redirect
+    a slash-prefixed prompt to a different command.
+    """
+    from specify_cli.integrations.opencode import OpencodeIntegration
+
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_OPENCODE_EXTRA_ARGS", "--command operator-override"
+    )
+    args = OpencodeIntegration().build_exec_args("/speckit body text")
+    # Prompt-derived "--command speckit" appears AFTER the
+    # operator-injected one, so a CLI that resolves repeated flags
+    # last-wins will honour Spec Kit's choice.
+    assert args == [
+        "opencode",
+        "run",
+        "--command",
+        "operator-override",
+        "--command",
+        "speckit",
+        "--format",
+        "json",
+        "body text",
+    ]
+
+
+def test_copilot_integration_honours_extra_args(monkeypatch):
+    from specify_cli.integrations.copilot import (
+        CopilotIntegration,
+        _copilot_executable,
+    )
+
+    # Disable --yolo so the argv shape stays deterministic.
+    monkeypatch.setenv("SPECKIT_COPILOT_ALLOW_ALL_TOOLS", "0")
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_COPILOT_EXTRA_ARGS", "--allow-tool 'shell(echo)'"
+    )
+    args = CopilotIntegration().build_exec_args("p")
+    # `_copilot_executable()` returns "copilot.cmd" on Windows and
+    # "copilot" elsewhere; the test must mirror that to stay portable.
+    assert args == [
+        _copilot_executable(),
+        "-p",
+        "p",
+        "--allow-tool",
+        "shell(echo)",
+        "--output-format",
+        "json",
+    ]
+
+
+# ---------------------------------------------------------------------------
+# `dispatch_command` end-to-end coverage
+#
+# Workflow execution calls `impl.dispatch_command(...)`, not
+# `build_exec_args` directly. `IntegrationBase.dispatch_command` delegates
+# to `build_exec_args` (so the override fixes above flow through), but
+# `CopilotIntegration` overrides `dispatch_command` and constructs
+# `cli_args` inline — the hook must be invoked there too or the env var
+# is silently ignored at workflow runtime. These tests monkeypatch
+# `subprocess.run` and assert the env-var args reach the executed argv.
+# ---------------------------------------------------------------------------
+
+
+class _RunCapture:
+    """Test double that captures argv passed to subprocess.run."""
+
+    def __init__(self):
+        self.captured_args: list[str] | None = None
+
+    def __call__(self, args, **kwargs):
+        self.captured_args = list(args)
+
+        class _Result:
+            returncode = 0
+            stdout = ""
+            stderr = ""
+
+        return _Result()
+
+
+def test_copilot_dispatch_command_includes_extra_args(monkeypatch):
+    """Locks the bypass fix: `CopilotIntegration.dispatch_command`
+    must honour `SPECKIT_INTEGRATION_COPILOT_EXTRA_ARGS`, not just `build_exec_args`.
+    """
+    import subprocess
+
+    from specify_cli.integrations.copilot import CopilotIntegration
+
+    capture = _RunCapture()
+    monkeypatch.setattr(subprocess, "run", capture)
+    monkeypatch.setenv("SPECKIT_COPILOT_ALLOW_ALL_TOOLS", "0")
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_COPILOT_EXTRA_ARGS", "--allow-tool 'shell(echo)'"
+    )
+
+    CopilotIntegration().dispatch_command(
+        "speckit.plan", args="body", stream=False
+    )
+
+    assert capture.captured_args is not None
+    # Hook inserted between `-p prompt` and the canonical Copilot flags.
+    p_idx = capture.captured_args.index("-p")
+    agent_idx = capture.captured_args.index("--agent")
+    extra_idx = capture.captured_args.index("--allow-tool")
+    assert p_idx < extra_idx < agent_idx
+    assert "shell(echo)" in capture.captured_args
+
+
+def test_codex_dispatch_command_includes_extra_args(monkeypatch):
+    """Lock the inherited `IntegrationBase.dispatch_command` path:
+    Codex (and by transitivity Devin, Opencode) flow through
+    `build_exec_args`, so the env var must reach argv at workflow
+    runtime.
+    """
+    import subprocess
+
+    from specify_cli.integrations.codex import CodexIntegration
+
+    capture = _RunCapture()
+    monkeypatch.setattr(subprocess, "run", capture)
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CODEX_EXTRA_ARGS", "--sandbox read-only")
+
+    CodexIntegration().dispatch_command(
+        "speckit.plan", args="body", stream=False
+    )
+
+    assert capture.captured_args is not None
+    assert "--sandbox" in capture.captured_args
+    assert "read-only" in capture.captured_args
+
+
+# ---------------------------------------------------------------------------
+# SPECKIT_INTEGRATION_<KEY>_EXECUTABLE tests
+#
+# The `_resolve_executable()` method on `IntegrationBase` checks
+# `SPECKIT_INTEGRATION_<KEY>_EXECUTABLE` and, when set, substitutes that
+# value for `self.key` as the first token in argv.  The tests below lock
+# the behaviour across shared and override integration paths:
+#   - the shared SkillsIntegration/MarkdownIntegration/TomlIntegration bases,
+#   - representative override integrations,
+#   - the hyphen→underscore key normalisation, and
+#   - whitespace/unset no-op guarantee.
+# ---------------------------------------------------------------------------
+
+
+def test_executable_env_var_unset_uses_key():
+    """Default: no override → executable is the integration key."""
+    args = _ClaudeStub().build_exec_args("p")
+    assert args[0] == "claude"
+
+
+def test_executable_env_var_replaces_first_argv_token(monkeypatch):
+    """Setting the env var substitutes the executable name in argv."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CLAUDE_EXECUTABLE", "/opt/claude/bin/claude")
+    args = _ClaudeStub().build_exec_args("hello")
+    assert args[0] == "/opt/claude/bin/claude"
+    assert args[1:] == ["-p", "hello", "--output-format", "json"]
+
+
+def test_executable_env_var_whitespace_only_falls_back_to_key(monkeypatch):
+    """Whitespace-only value is treated as unset → falls back to self.key."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CLAUDE_EXECUTABLE", "   ")
+    args = _ClaudeStub().build_exec_args("p")
+    assert args[0] == "claude"
+
+
+def test_executable_env_var_key_normalization_hyphen_to_underscore(monkeypatch):
+    """`kiro-cli` key maps to `SPECKIT_INTEGRATION_KIRO_CLI_EXECUTABLE`."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_KIRO_CLI_EXECUTABLE", "/usr/local/bin/kiro-cli")
+    args = _KiroCliStub().build_exec_args("p")
+    assert args[0] == "/usr/local/bin/kiro-cli"
+
+
+def test_executable_env_var_other_integration_ignored(monkeypatch):
+    """`SPECKIT_INTEGRATION_GEMINI_EXECUTABLE` must NOT affect Claude."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_GEMINI_EXECUTABLE", "/custom/gemini")
+    args = _ClaudeStub().build_exec_args("p")
+    assert args[0] == "claude"
+
+
+def test_executable_env_var_markdown_integration(monkeypatch):
+    """MarkdownIntegration base honours the executable env var."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_MD_AGENT_EXECUTABLE", "/custom/md-agent")
+    args = _MarkdownAgentStub().build_exec_args("p")
+    assert args[0] == "/custom/md-agent"
+
+
+def test_executable_env_var_toml_integration(monkeypatch):
+    """TomlIntegration base honours the executable env var."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_TOML_AGENT_EXECUTABLE", "/custom/toml-agent")
+    args = _TomlAgentStub().build_exec_args("p")
+    assert args[0] == "/custom/toml-agent"
+
+
+def test_executable_env_var_requires_cli_false_returns_none(monkeypatch):
+    """`requires_cli: False` still returns None even when executable is set."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_NO_CLI_EXECUTABLE", "/custom/no-cli")
+    assert _NoCliStub().build_exec_args("p") is None
+
+
+def test_executable_env_var_codex_integration(monkeypatch):
+    """CodexIntegration honours the executable env var."""
+    from specify_cli.integrations.codex import CodexIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CODEX_EXECUTABLE", "/opt/codex")
+    args = CodexIntegration().build_exec_args("p")
+    assert args[0] == "/opt/codex"
+    assert args[1] == "exec"
+
+
+def test_executable_env_var_devin_integration(monkeypatch):
+    """DevinIntegration honours the executable env var."""
+    from specify_cli.integrations.devin import DevinIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_DEVIN_EXECUTABLE", "/opt/devin")
+    args = DevinIntegration().build_exec_args("p")
+    assert args[0] == "/opt/devin"
+
+
+def test_executable_env_var_opencode_integration(monkeypatch):
+    """OpencodeIntegration honours the executable env var."""
+    from specify_cli.integrations.opencode import OpencodeIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_OPENCODE_EXECUTABLE", "/opt/opencode")
+    args = OpencodeIntegration().build_exec_args("p")
+    assert args[0] == "/opt/opencode"
+    assert args[1] == "run"
+
+
+def test_executable_env_var_copilot_integration(monkeypatch):
+    """CopilotIntegration honours the executable env var, overriding the
+    platform-specific default from `_copilot_executable()`."""
+    from specify_cli.integrations.copilot import CopilotIntegration
+
+    monkeypatch.setenv("SPECKIT_INTEGRATION_COPILOT_EXECUTABLE", "/opt/copilot")
+    monkeypatch.setenv("SPECKIT_COPILOT_ALLOW_ALL_TOOLS", "0")
+    args = CopilotIntegration().build_exec_args("p")
+    assert args[0] == "/opt/copilot"
+
+
+def test_executable_env_var_copilot_unset_uses_platform_default(monkeypatch):
+    """When `SPECKIT_INTEGRATION_COPILOT_EXECUTABLE` is unset, Copilot
+    falls back to the platform-specific default from `_copilot_executable()`."""
+    from specify_cli.integrations.copilot import CopilotIntegration, _copilot_executable
+
+    monkeypatch.setenv("SPECKIT_COPILOT_ALLOW_ALL_TOOLS", "0")
+    args = CopilotIntegration().build_exec_args("p")
+    assert args[0] == _copilot_executable()
+
+
+def test_executable_env_var_copilot_dispatch_command(monkeypatch):
+    """CopilotIntegration.dispatch_command honours the executable env var."""
+    import subprocess
+
+    from specify_cli.integrations.copilot import CopilotIntegration
+
+    capture = _RunCapture()
+    monkeypatch.setattr(subprocess, "run", capture)
+    monkeypatch.setenv("SPECKIT_INTEGRATION_COPILOT_EXECUTABLE", "/opt/copilot")
+    monkeypatch.setenv("SPECKIT_COPILOT_ALLOW_ALL_TOOLS", "0")
+
+    CopilotIntegration().dispatch_command("speckit.plan", args="body", stream=False)
+
+    assert capture.captured_args is not None
+    assert capture.captured_args[0] == "/opt/copilot"
+
+
+def test_executable_and_extra_args_both_honoured(monkeypatch):
+    """Both the executable override and extra args env vars can be set
+    simultaneously — they are independent hooks."""
+    monkeypatch.setenv("SPECKIT_INTEGRATION_CLAUDE_EXECUTABLE", "/opt/claude")
+    monkeypatch.setenv(
+        "SPECKIT_INTEGRATION_CLAUDE_EXTRA_ARGS", "--dangerously-skip-permissions"
+    )
+    args = _ClaudeStub().build_exec_args("hello", model="sonnet")
+    assert args == [
+        "/opt/claude",
+        "-p",
+        "hello",
+        "--dangerously-skip-permissions",
+        "--model",
+        "sonnet",
+        "--output-format",
+        "json",
+    ]

tests/integrations/test_integration_agy.py

@@ -1,5 +1,7 @@
 """Tests for AgyIntegration (Antigravity)."""
 
+from specify_cli.integrations import get_integration
+
 from .test_integration_base_skills import SkillsIntegrationTests
 
 
@@ -12,10 +14,21 @@ class TestAgyIntegration(SkillsIntegrationTests):
 
     def test_options_include_skills_flag(self):
         """Override inherited test: AgyIntegration should not expose a --skills flag because .agents/ is its only layout."""
-        from specify_cli.integrations import get_integration
         i = get_integration(self.KEY)
         skills_opts = [o for o in i.options() if o.name == "--skills"]
         assert len(skills_opts) == 0
+
+    def test_requires_cli_is_true(self):
+        """agy is a CLI tool; requires_cli must be True."""
+        i = get_integration(self.KEY)
+        assert i.config["requires_cli"] is True
+
+    def test_install_url_is_set(self):
+        """install_url must point to the official installation page."""
+        i = get_integration(self.KEY)
+        assert i.config["install_url"] == "https://antigravity.google/"
+
+
 class TestAgyAutoPromote:
     """--ai agy auto-promotes to integration path."""
 
@@ -26,7 +39,7 @@ class TestAgyAutoPromote:
 
         runner = CliRunner()
         target = tmp_path / "test-proj"
-        result = runner.invoke(app, ["init", str(target), "--ai", "agy", "--no-git", "--script", "sh"])
+        result = runner.invoke(app, ["init", str(target), "--ai", "agy", "--no-git", "--script", "sh", "--ignore-agent-tools"])
 
         assert result.exit_code == 0, f"init --ai agy failed: {result.output}"
         assert (target / ".agents" / "skills" / "speckit-plan" / "SKILL.md").exists()
@@ -36,10 +49,87 @@ class TestAgyAutoPromote:
         from typer.testing import CliRunner
         from specify_cli import app
 
-        # Click >= 8.2 separates stdout and stderr natively, mix_stderr is removed
+        # Click >= 8.2 separates stdout and stderr natively
         runner = CliRunner()
         target = tmp_path / "test-proj2"
-        result = runner.invoke(app, ["init", str(target), "--ai", "agy", "--no-git", "--script", "sh"])
+        result = runner.invoke(app, ["init", str(target), "--ai", "agy", "--no-git", "--script", "sh", "--ignore-agent-tools"])
 
         assert result.exit_code == 0
         assert "Warning: The .agents/ layout requires Antigravity v1.20.5 or newer" in result.stderr
+
+
+class TestAgyBuildExecArgs:
+    """agy non-interactive execution argument building."""
+
+    def test_build_exec_args_returns_print_command(self):
+        """build_exec_args should return ['agy', '--print', prompt]."""
+        from specify_cli.integrations import get_integration
+        i = get_integration("agy")
+        result = i.build_exec_args("describe my feature")
+        assert result == ["agy", "--print", "describe my feature"]
+
+    def test_build_exec_args_ignores_model(self):
+        """agy does not support --model; model param must be ignored."""
+        from specify_cli.integrations import get_integration
+        i = get_integration("agy")
+        result = i.build_exec_args("my prompt", model="gemini-pro")
+        assert result == ["agy", "--print", "my prompt"]
+
+    def test_build_exec_args_ignores_output_json(self):
+        """agy does not support JSON output; output_json param must be ignored."""
+        from specify_cli.integrations import get_integration
+        i = get_integration("agy")
+        result = i.build_exec_args("my prompt", output_json=False)
+        assert result == ["agy", "--print", "my prompt"]
+
+
+class TestAgyHookCommandNote:
+    """Verify dot-to-hyphen normalization note is injected into hook sections."""
+
+    def test_hook_note_injected_in_skills_with_hooks(self, tmp_path):
+        """Skills with hook sections should contain the normalization note."""
+        from specify_cli.integrations import get_integration
+        from specify_cli.integrations.manifest import IntegrationManifest
+
+        i = get_integration("agy")
+        m = IntegrationManifest("agy", tmp_path)
+        i.setup(tmp_path, m, script_type="sh")
+        specify_skill = tmp_path / ".agents/skills/speckit-specify/SKILL.md"
+        assert specify_skill.exists()
+        content = specify_skill.read_text(encoding="utf-8")
+        assert "replace dots" in content, (
+            "speckit-specify should have dot-to-hyphen hook note"
+        )
+
+    def test_hook_note_not_in_skills_without_hooks(self):
+        """Skills without hook sections should not get the note."""
+        from specify_cli.integrations.agy import AgyIntegration
+
+        content = "---\nname: test\ndescription: test\n---\n\nNo hooks here.\n"
+        result = AgyIntegration._inject_hook_command_note(content)
+        assert "replace dots" not in result
+
+    def test_hook_note_idempotent(self):
+        """Injecting the note twice must not duplicate it."""
+        from specify_cli.integrations.agy import AgyIntegration
+
+        content = (
+            "---\nname: test\n---\n\n"
+            "- For each executable hook, output the following based on its flag:\n"
+        )
+        once = AgyIntegration._inject_hook_command_note(content)
+        twice = AgyIntegration._inject_hook_command_note(once)
+        assert once == twice, "Hook note injection should be idempotent"
+
+    def test_hook_note_preserves_indentation(self):
+        """The injected note must match the indentation of the target line."""
+        from specify_cli.integrations.agy import AgyIntegration
+
+        content = (
+            "---\nname: test\n---\n\n"
+            "   - For each executable hook, output the following\n"
+        )
+        result = AgyIntegration._inject_hook_command_note(content)
+        lines = result.splitlines()
+        note_line = [l for l in lines if "replace dots" in l][0]
+        assert note_line.startswith("   "), "Note should preserve indentation"

tests/integrations/test_integration_base_skills.py

@@ -176,6 +176,39 @@ class SkillsIntegrationTests:
                 f"skills agents must use /speckit-<name>"
             )
 
+    def test_hook_sections_explain_dotted_command_conversion(self, tmp_path):
+        """Generated skills with hook sections must explain dotted command conversion."""
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        i.setup(tmp_path, m)
+        specify_skill = i.skills_dest(tmp_path) / "speckit-specify" / "SKILL.md"
+        assert specify_skill.exists()
+        content = specify_skill.read_text(encoding="utf-8")
+        assert "replace dots" in content, (
+            "speckit-specify should explain dotted hook command conversion"
+        )
+        assert content.count("replace dots") == content.count(
+            "- For each executable hook, output the following"
+        )
+
+    def test_hook_note_injected_for_each_instruction_independently(self):
+        """Existing hook notes should not suppress later missing notes."""
+        content = (
+            "---\n"
+            "name: test\n"
+            "---\n\n"
+            "- When constructing slash commands from hook command names, "
+            "replace dots (`.`) with hyphens (`-`). "
+            "For example, `speckit.git.commit` → `/speckit-git-commit`.\n"
+            "- For each executable hook, output the following first block:\n"
+            "\n"
+            "- For each executable hook, output the following second block:\n"
+        )
+
+        result = SkillsIntegration._inject_hook_command_note(content)
+
+        assert result.count("replace dots (`.`) with hyphens") == 2
+
     def test_skill_body_has_content(self, tmp_path):
         """Each SKILL.md body should contain template content after the frontmatter."""
         i = get_integration(self.KEY)

tests/integrations/test_integration_claude.py

@@ -8,7 +8,7 @@ from unittest.mock import patch
 import yaml
 
 from specify_cli.integrations import INTEGRATION_REGISTRY, get_integration
-from specify_cli.integrations.base import IntegrationBase
+from specify_cli.integrations.base import IntegrationBase, SkillsIntegration
 from specify_cli.integrations.claude import ARGUMENT_HINTS
 from specify_cli.integrations.manifest import IntegrationManifest
 
@@ -197,8 +197,8 @@ class TestClaudeIntegration:
             os.chdir(project)
             runner = CliRunner()
             with (
-                patch("specify_cli._stdin_is_interactive", return_value=True),
-                patch("specify_cli.select_with_arrows", return_value="claude"),
+                patch("specify_cli.commands.init._stdin_is_interactive", return_value=True),
+                patch("specify_cli.commands.init.select_with_arrows", return_value="claude"),
             ):
                 result = runner.invoke(
                     app,
@@ -487,8 +487,8 @@ class TestClaudeDisableModelInvocation:
         assert "disable-model-invocation" not in fm
         assert "user-invocable" not in fm
 
-    def test_skills_default_post_process_is_identity(self, tmp_path):
-        """SkillsIntegration agents without an override leave content unchanged."""
+    def test_skills_default_post_process_preserves_content_without_hooks(self, tmp_path):
+        """SkillsIntegration agents without an override preserve non-hook content."""
         # ``agy`` is a plain SkillsIntegration with no post-process override,
         # so it stands in for the base-class default behavior.
         agy = get_integration("agy")
@@ -505,7 +505,7 @@ class TestClaudeHookCommandNote:
         """Skills that have hook sections should get the normalization note."""
         i = get_integration("claude")
         m = IntegrationManifest("claude", tmp_path)
-        created = i.setup(tmp_path, m, script_type="sh")
+        i.setup(tmp_path, m, script_type="sh")
         specify_skill = tmp_path / ".claude/skills/speckit-specify/SKILL.md"
         assert specify_skill.exists()
         content = specify_skill.read_text(encoding="utf-8")
@@ -516,35 +516,54 @@ class TestClaudeHookCommandNote:
 
     def test_hook_note_not_in_skills_without_hooks(self, tmp_path):
         """Skills without hook sections should not get the note."""
-        from specify_cli.integrations.claude import ClaudeIntegration
-
         content = "---\nname: test\ndescription: test\n---\n\nNo hooks here.\n"
-        result = ClaudeIntegration._inject_hook_command_note(content)
+        result = SkillsIntegration._inject_hook_command_note(content)
         assert "replace dots" not in result
 
     def test_hook_note_idempotent(self, tmp_path):
         """Injecting the note twice should not duplicate it."""
-        from specify_cli.integrations.claude import ClaudeIntegration
-
         content = (
             "---\nname: test\n---\n\n"
             "- For each executable hook, output the following based on its flag:\n"
         )
-        once = ClaudeIntegration._inject_hook_command_note(content)
-        twice = ClaudeIntegration._inject_hook_command_note(once)
+        once = SkillsIntegration._inject_hook_command_note(content)
+        twice = SkillsIntegration._inject_hook_command_note(once)
         assert once == twice, "Hook note injection should be idempotent"
 
+    def test_hook_note_fills_missing_repeated_instructions(self, tmp_path):
+        """Already-noted hook sections should not suppress later sections."""
+        from specify_cli.integrations.base import _HOOK_COMMAND_NOTE
+
+        content = (
+            "---\nname: test\n---\n\n"
+            f"{_HOOK_COMMAND_NOTE}"
+            "- For each executable hook, output the following based on its flag:\n"
+            "\n"
+            "  - For each executable hook, output the following based on its flag:\n"
+        )
+        result = SkillsIntegration._inject_hook_command_note(content)
+        assert result.count("replace dots (`.`) with hyphens") == 2
+
+    def test_hook_note_not_suppressed_by_unrelated_phrase(self, tmp_path):
+        """Unrelated text should not trip the hook-note idempotence guard."""
+        content = (
+            "---\nname: test\n---\n\n"
+            "This paragraph says replace dots in a different context.\n"
+            "- For each executable hook, output the following based on its flag:\n"
+        )
+        result = SkillsIntegration._inject_hook_command_note(content)
+        assert "This paragraph says replace dots in a different context." in result
+        assert result.count("replace dots (`.`) with hyphens") == 1
+
     def test_hook_note_preserves_indentation(self, tmp_path):
         """The injected note should match the indentation of the target line."""
-        from specify_cli.integrations.claude import ClaudeIntegration
-
         content = (
             "---\nname: test\n---\n\n"
             "   - For each executable hook, output the following\n"
         )
-        result = ClaudeIntegration._inject_hook_command_note(content)
+        result = SkillsIntegration._inject_hook_command_note(content)
         lines = result.splitlines()
-        note_line = [l for l in lines if "replace dots" in l][0]
+        note_line = [line for line in lines if "replace dots" in line][0]
         assert note_line.startswith("   "), "Note should preserve indentation"
 
     def test_post_process_injects_all_claude_flags(self):

tests/integrations/test_integration_codex.py

@@ -71,6 +71,34 @@ class TestCodexHookCommandNote:
         twice = CodexIntegration._inject_hook_command_note(once)
         assert once == twice, "Hook note injection should be idempotent"
 
+    def test_hook_note_fills_missing_repeated_instructions(self):
+        """Already-noted hook sections should not suppress later sections."""
+        from specify_cli.integrations.base import _HOOK_COMMAND_NOTE
+        from specify_cli.integrations.codex import CodexIntegration
+
+        content = (
+            "---\nname: test\n---\n\n"
+            f"{_HOOK_COMMAND_NOTE}"
+            "- For each executable hook, output the following based on its flag:\n"
+            "\n"
+            "  - For each executable hook, output the following based on its flag:\n"
+        )
+        result = CodexIntegration._inject_hook_command_note(content)
+        assert result.count("replace dots (`.`) with hyphens") == 2
+
+    def test_hook_note_not_suppressed_by_unrelated_phrase(self):
+        """Unrelated text should not trip the hook-note idempotence guard."""
+        from specify_cli.integrations.codex import CodexIntegration
+
+        content = (
+            "---\nname: test\n---\n\n"
+            "This paragraph says replace dots in a different context.\n"
+            "- For each executable hook, output the following based on its flag:\n"
+        )
+        result = CodexIntegration._inject_hook_command_note(content)
+        assert "This paragraph says replace dots in a different context." in result
+        assert result.count("replace dots (`.`) with hyphens") == 1
+
     def test_hook_note_preserves_indentation(self):
         """The injected note should match the indentation of the target line."""
         from specify_cli.integrations.codex import CodexIntegration
@@ -81,7 +109,7 @@ class TestCodexHookCommandNote:
         )
         result = CodexIntegration._inject_hook_command_note(content)
         lines = result.splitlines()
-        note_line = [l for l in lines if "replace dots" in l][0]
+        note_line = [line for line in lines if "replace dots" in line][0]
         assert note_line.startswith("   "), "Note should preserve indentation"
 
     def test_hook_note_when_instruction_is_final_line_without_newline(self):
@@ -102,11 +130,11 @@ class TestCodexHookCommandNote:
         result = CodexIntegration._inject_hook_command_note(content)
         lines = result.splitlines()
         note_line_idx = next(
-            i for i, l in enumerate(lines) if "replace dots" in l
+            i for i, line in enumerate(lines) if "replace dots" in line
         )
         instruction_line_idx = next(
-            i for i, l in enumerate(lines)
-            if l.lstrip().startswith("- For each executable hook")
+            i for i, line in enumerate(lines)
+            if line.lstrip().startswith("- For each executable hook")
         )
         assert note_line_idx < instruction_line_idx, (
             "Note must appear before the instruction"

tests/integrations/test_integration_copilot.py

@@ -404,6 +404,20 @@ class TestCopilotSkillsMode:
         updated = copilot.post_process_skill_content(content)
         assert "mode: speckit.plan" in updated
 
+    def test_post_process_skill_content_injects_hook_note(self):
+        """post_process_skill_content() should inject shared hook guidance."""
+        copilot = self._make_copilot()
+        content = (
+            "---\n"
+            'name: "speckit-specify"\n'
+            'description: "Specify workflow"\n'
+            "---\n"
+            "\n- For each executable hook, output the following\n"
+        )
+        updated = copilot.post_process_skill_content(content)
+        assert "replace dots" in updated
+        assert "mode: speckit.specify" in updated
+
     def test_post_process_idempotent(self):
         """post_process_skill_content() must be idempotent."""
         copilot = self._make_copilot()
@@ -434,6 +448,14 @@ class TestCopilotSkillsMode:
             stem = skill_dir_name.removeprefix("speckit-")
             assert fm["mode"] == f"speckit.{stem}"
 
+    def test_skills_hook_sections_explain_dotted_command_conversion(self, tmp_path):
+        """Generated skills with hook sections should include shared hook guidance."""
+        copilot = self._make_copilot()
+        self._setup_skills(copilot, tmp_path)
+        specify_skill = tmp_path / ".github" / "skills" / "speckit-specify" / "SKILL.md"
+        content = specify_skill.read_text(encoding="utf-8")
+        assert "replace dots" in content
+
     # -- Template processing ----------------------------------------------
 
     def test_skills_templates_are_processed(self, tmp_path):
@@ -724,4 +746,4 @@ class TestCopilotSkillsMode:
         # Must NOT show the dotted /speckit.plan form
         assert "/speckit.plan" not in result.output, (
             f"Should not show /speckit.plan in skills mode:\n{result.output}"
-        )
+        )

tests/integrations/test_integration_hermes.py

@@ -0,0 +1,347 @@
+"""Tests for HermesIntegration.
+
+Hermes is special among SkillsIntegration subclasses: it writes skills
+to ``~/.hermes/skills/`` (global) rather than the project-local
+``.hermes/skills/`` directory.  A project-local marker (empty directory)
+is created so extension commands (e.g. git) can detect Hermes.
+
+All tests that touch ``~/.hermes/`` use ``monkeypatch`` to isolate
+``Path.home()`` to a temp directory so the test suite is hermetic and
+non-destructive to a developer's real Hermes installation.
+"""
+
+from pathlib import Path
+
+from specify_cli.integrations import get_integration
+from specify_cli.integrations.manifest import IntegrationManifest
+
+from .test_integration_base_skills import SkillsIntegrationTests
+
+
+def _fake_home(tmp_path: Path) -> Path:
+    """Create and return an isolated home directory under *tmp_path*."""
+    home = tmp_path / "home"
+    home.mkdir(exist_ok=True)
+    return home
+
+
+class TestHermesIntegration(SkillsIntegrationTests):
+    KEY = "hermes"
+    FOLDER = ".hermes/"
+    COMMANDS_SUBDIR = "skills"
+    REGISTRAR_DIR = "~/.hermes/skills"
+    CONTEXT_FILE = "AGENTS.md"
+
+    # -- Hermes-specific setup: skills go to ~/.hermes/skills/ -------------
+
+    def test_setup_writes_to_global_skills_dir(self, tmp_path, monkeypatch):
+        """Skills are written to ~/.hermes/skills/, not project-local."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.setup(tmp_path, m)
+        skill_files = [f for f in created if "scripts" not in f.parts]
+
+        assert len(skill_files) > 0, "No skill files were created"
+        for f in skill_files:
+            # Every skill file should be under ~/.hermes/skills/speckit-*/
+            expected_prefix = str(home / ".hermes" / "skills")
+            assert str(f).startswith(expected_prefix), (
+                f"{f} is not under ~/.hermes/skills/"
+            )
+
+    def test_local_marker_dir_created(self, tmp_path, monkeypatch):
+        """Project-local .hermes/skills/ should exist but be empty."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        i.setup(tmp_path, m)
+        marker = tmp_path / ".hermes" / "skills"
+        assert marker.is_dir(), "Marker directory was not created"
+        # Should be empty (no SKILL.md files)
+        children = list(marker.iterdir())
+        assert children == [], f"Marker directory should be empty, got: {children}"
+
+    # -- Override shared tests that assume project-local skills ------------
+
+    def test_setup_writes_to_correct_directory(self, tmp_path, monkeypatch):
+        """Override: Hermes writes to global, not project-local."""
+        self.test_setup_writes_to_global_skills_dir(tmp_path, monkeypatch)
+
+    def test_plan_references_correct_context_file(self, tmp_path, monkeypatch):
+        """Plan skill goes to global dir, but we check it still references AGENTS.md."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        if not i.context_file:
+            return
+        m = IntegrationManifest(self.KEY, tmp_path)
+        i.setup(tmp_path, m)
+        # Find the plan skill in global ~/.hermes/skills/
+        plan_file = home / ".hermes" / "skills" / "speckit-plan" / "SKILL.md"
+        assert plan_file.exists(), f"Plan skill {plan_file} not created globally"
+        content = plan_file.read_text(encoding="utf-8")
+        assert i.context_file in content, (
+            f"Plan skill should reference {i.context_file!r} but it was not found"
+        )
+        assert "__CONTEXT_FILE__" not in content, (
+            "Plan skill has unprocessed __CONTEXT_FILE__ placeholder"
+        )
+
+    def test_all_files_tracked_in_manifest(self, tmp_path, monkeypatch):
+        """Override: Hermes does not track skills in the project manifest
+        since they live globally.  Only project-local files (scripts,
+        templates, context) are tracked."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.setup(tmp_path, m)
+        for f in created:
+            # Global files (in ~/.hermes/) are not tracked in manifest
+            if str(f).startswith(str(home)):
+                continue
+            rel = f.resolve().relative_to(tmp_path.resolve()).as_posix()
+            assert rel in m.files, f"{rel} not tracked in manifest"
+
+    def test_install_uninstall_roundtrip(self, tmp_path, monkeypatch):
+        """Override: Hermes uninstall removes global skills + local marker."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.install(tmp_path, m)
+        assert len(created) > 0
+        m.save()
+        # All SKILL.md files should exist globally
+        for f in created:
+            if "SKILL.md" in str(f):
+                assert f.exists(), f"{f} does not exist"
+        # Global skills are removed on teardown without needing force
+        removed, skipped = i.teardown(tmp_path, m, force=False)
+        for f in created:
+            if "SKILL.md" in str(f):
+                assert not f.exists(), f"{f} should have been removed"
+        # Local marker should be gone
+        assert not (tmp_path / ".hermes" / "skills").exists()
+
+    def test_modified_file_survives_uninstall(self, tmp_path, monkeypatch):
+        """Override: Hermes global skills are ALWAYS removed on uninstall
+        (they live outside the project root and aren't hash-tracked in the
+        manifest), so a modified global skill is still removed — matching
+        the standard behaviour where all integration files are cleaned up."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.install(tmp_path, m)
+        m.save()
+        # Pick a global skill file
+        skill_files = [f for f in created if "SKILL.md" in str(f)]
+        assert len(skill_files) > 0
+        modified_file = skill_files[0]
+        modified_file.write_text("user modified this", encoding="utf-8")
+        removed, skipped = i.uninstall(tmp_path, m)
+        assert not modified_file.exists(), (
+            "Modified global skill should be removed on teardown (standard behaviour)"
+        )
+
+    def test_modified_global_skill_removed_on_teardown(self, tmp_path, monkeypatch):
+        """Override: Hermes global skills are removed on uninstall regardless
+        of the force flag, matching standard integration behaviour."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.install(tmp_path, m)
+        m.save()
+        # Pick a global skill file
+        skill_files = [f for f in created if "SKILL.md" in str(f)]
+        assert len(skill_files) > 0
+        modified_file = skill_files[0]
+        modified_file.write_text("user modified this", encoding="utf-8")
+        # Global skills are removed on teardown regardless of force flag
+        removed, skipped = i.teardown(tmp_path, m, force=False)
+        assert not modified_file.exists(), (
+            "Modified global skill should be removed on teardown (standard behaviour)"
+        )
+
+    def test_pre_existing_skills_not_removed(self, tmp_path, monkeypatch):
+        """Pre-existing non-speckit global skills should survive Hermes uninstall."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        # Create a foreign skill in the global dir first
+        global_skills_dir = i._hermes_home_skills_dir()
+        foreign_dir = global_skills_dir / "other-tool"
+        foreign_dir.mkdir(parents=True, exist_ok=True)
+        (foreign_dir / "SKILL.md").write_text("# Foreign skill\n")
+
+        m = IntegrationManifest(self.KEY, tmp_path)
+        i.setup(tmp_path, m)
+
+        # Run teardown to verify foreign skill survives uninstall
+        i.teardown(tmp_path, m)
+
+        assert (foreign_dir / "SKILL.md").exists(), (
+            "Foreign skill was removed by teardown"
+        )
+
+    def test_hook_sections_explain_dotted_command_conversion(self, tmp_path, monkeypatch):
+        """Override: Hermes skills live in global ~/.hermes/skills/."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        i.setup(tmp_path, m)
+        specify_skill = home / ".hermes" / "skills" / "speckit-specify" / "SKILL.md"
+        assert specify_skill.exists()
+        content = specify_skill.read_text(encoding="utf-8")
+        assert "replace dots" in content, (
+            "speckit-specify should explain dotted hook command conversion"
+        )
+        assert content.count("replace dots") == content.count(
+            "- For each executable hook, output the following"
+        )
+
+    def test_complete_file_inventory_sh(self, tmp_path, monkeypatch):
+        """Override: Hermes init produces no local SKILL.md files,
+        only the empty .hermes/skills/ marker."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        from typer.testing import CliRunner
+        from specify_cli import app
+
+        project = tmp_path / f"inventory-sh-{self.KEY}"
+        project.mkdir()
+        old_cwd = Path.cwd()
+        import os
+        try:
+            os.chdir(project)
+            result = CliRunner().invoke(app, [
+                "init", "--here", "--integration", self.KEY,
+                "--script", "sh", "--no-git", "--ignore-agent-tools",
+            ], catch_exceptions=False)
+        finally:
+            os.chdir(old_cwd)
+        assert result.exit_code == 0, f"init failed: {result.output}"
+        actual = sorted(
+            p.relative_to(project).as_posix()
+            for p in project.rglob("*") if p.is_file()
+        )
+        # Ensure no .hermes/skills/speckit-*/SKILL.md in project dir
+        hermes_skill_files = [f for f in actual if f.startswith(".hermes/skills/speckit-")]
+        assert hermes_skill_files == [], (
+            f"Expected no local SKILL.md files, found: {hermes_skill_files}"
+        )
+        # Ensure the marker exists (empty dir won't appear in file listing)
+        assert (project / ".hermes" / "skills").is_dir()
+
+    def test_complete_file_inventory_ps(self, tmp_path, monkeypatch):
+        """Override: Same as sh variant but for PowerShell script type."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        from typer.testing import CliRunner
+        from specify_cli import app
+
+        project = tmp_path / f"inventory-ps-{self.KEY}"
+        project.mkdir()
+        old_cwd = Path.cwd()
+        import os
+        try:
+            os.chdir(project)
+            result = CliRunner().invoke(app, [
+                "init", "--here", "--integration", self.KEY,
+                "--script", "ps", "--no-git", "--ignore-agent-tools",
+            ], catch_exceptions=False)
+        finally:
+            os.chdir(old_cwd)
+        assert result.exit_code == 0, f"init failed: {result.output}"
+        actual = sorted(
+            p.relative_to(project).as_posix()
+            for p in project.rglob("*") if p.is_file()
+        )
+        hermes_skill_files = [f for f in actual if f.startswith(".hermes/skills/speckit-")]
+        assert hermes_skill_files == [], (
+            f"Expected no local SKILL.md files, found: {hermes_skill_files}"
+        )
+        assert (project / ".hermes" / "skills").is_dir()
+
+    def test_install_uninstall_cleanup(self, tmp_path, monkeypatch):
+        """Verify global skills are cleaned and local marker is removed."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        i = get_integration(self.KEY)
+        m = IntegrationManifest(self.KEY, tmp_path)
+        created = i.setup(tmp_path, m)
+
+        # Verify global skills exist
+        global_skills = [
+            f for f in created
+            if "SKILL.md" in str(f)
+            and str(f).startswith(str(home / ".hermes"))
+        ]
+        assert len(global_skills) > 0
+        for f in global_skills:
+            assert f.exists()
+
+        # Verify local marker exists
+        assert (tmp_path / ".hermes" / "skills").is_dir()
+
+        # Teardown — global skills removed without needing force=True
+        removed, skipped = i.teardown(tmp_path, m, force=False)
+
+        # Global skills removed
+        for f in global_skills:
+            assert not f.exists(), f"{f} should have been removed"
+
+        # Local marker removed
+        assert not (tmp_path / ".hermes" / "skills").exists(), (
+            "Local marker should be removed on teardown"
+        )
+
+
+class TestHermesAutoPromote:
+    """--ai hermes auto-promotes to integration path."""
+
+    def test_ai_hermes_without_ai_skills_auto_promotes(self, tmp_path, monkeypatch):
+        """--ai hermes should work the same as --integration hermes,
+        creating global skills and a local marker."""
+        home = _fake_home(tmp_path)
+        monkeypatch.setattr(Path, "home", lambda: home)
+
+        from typer.testing import CliRunner
+        from specify_cli import app
+
+        runner = CliRunner()
+        target = tmp_path / "test-proj"
+        result = runner.invoke(app, [
+            "init", str(target),
+            "--ai", "hermes",
+            "--no-git",
+            "--ignore-agent-tools",
+            "--script", "sh",
+        ])
+
+        assert result.exit_code == 0, f"init --ai hermes failed: {result.output}"
+        # Skills should be in global ~/.hermes/skills/
+        assert (home / ".hermes" / "skills" / "speckit-plan" / "SKILL.md").exists()
+        # Local marker should exist
+        assert (target / ".hermes" / "skills").is_dir()
+        # No SKILL.md files in project-local dir
+        local_skills = list((target / ".hermes" / "skills").iterdir())
+        assert local_skills == [], f"Local skills dir should be empty, got: {local_skills}"

tests/integrations/test_integration_subcommand.py

@@ -7,6 +7,7 @@ import pytest
 from typer.testing import CliRunner
 
 from specify_cli import app
+from tests.conftest import strip_ansi
 
 
 runner = CliRunner()
@@ -49,7 +50,8 @@ def _write_invalid_manifest(project, key):
 
 
 def _integration_list_row_cells(output: str, key: str) -> list[str]:
-    row = next(line for line in output.splitlines() if line.startswith(f"│ {key}"))
+    plain = strip_ansi(output)
+    row = next(line for line in plain.splitlines() if line.startswith(f"│ {key}"))
     return [cell.strip() for cell in row.split("│")[1:-1]]
 
 
@@ -160,8 +162,9 @@ class TestIntegrationInstall:
         finally:
             os.chdir(old_cwd)
         assert result.exit_code == 0
-        assert "already installed" in result.output
-        normalized = " ".join(result.output.split())
+        plain = strip_ansi(result.output)
+        assert "already installed" in plain
+        normalized = " ".join(plain.split())
         assert "specify integration upgrade copilot" in normalized
         assert "already the default integration" in normalized
         assert "No files were changed" in normalized
@@ -197,9 +200,10 @@ class TestIntegrationInstall:
         finally:
             os.chdir(old_cwd)
         assert result.exit_code != 0
-        assert "Installed integrations: copilot" in result.output
-        assert "Default integration: copilot" in result.output
-        normalized = " ".join(result.output.split())
+        plain = strip_ansi(result.output)
+        assert "Installed integrations: copilot" in plain
+        assert "Default integration: copilot" in plain
+        normalized = " ".join(plain.split())
         assert "To replace the default integration" in normalized
         assert "specify integration switch claude" in normalized
         assert "To install 'claude' alongside" in normalized
@@ -230,6 +234,29 @@ class TestIntegrationInstall:
         assert (project / ".claude" / "skills" / "speckit-plan" / "SKILL.md").exists()
         assert (project / ".agents" / "skills" / "speckit-plan" / "SKILL.md").exists()
 
+    def test_install_non_default_refreshes_init_options_version_only(self, tmp_path, monkeypatch):
+        project = _init_project(tmp_path, "claude")
+        init_options = project / ".specify" / "init-options.json"
+        opts = json.loads(init_options.read_text(encoding="utf-8"))
+        opts["speckit_version"] = "0.6.1"
+        init_options.write_text(json.dumps(opts), encoding="utf-8")
+
+        import specify_cli
+
+        monkeypatch.setattr(specify_cli, "get_speckit_version", lambda: "0.8.11")
+
+        result = _run_in_project(project, [
+            "integration", "install", "codex",
+            "--script", "sh",
+        ])
+
+        assert result.exit_code == 0, result.output
+        updated = json.loads(init_options.read_text(encoding="utf-8"))
+        assert updated["speckit_version"] == "0.8.11"
+        assert updated["integration"] == "claude"
+        assert updated["ai"] == "claude"
+        assert updated["context_file"] == "CLAUDE.md"
+
     def test_install_additional_preserves_shared_manifest(self, tmp_path):
         project = _init_project(tmp_path, "claude")
         shared_manifest = project / ".specify" / "integrations" / "speckit.manifest.json"
@@ -286,9 +313,10 @@ class TestIntegrationInstall:
         finally:
             os.chdir(old_cwd)
         assert result.exit_code != 0
-        assert "Installed integrations: copilot" in result.output
-        assert "multi-install safe" in result.output
-        normalized = " ".join(result.output.split())
+        plain = strip_ansi(result.output)
+        assert "Installed integrations: copilot" in plain
+        assert "multi-install safe" in plain
+        normalized = " ".join(plain.split())
         assert "To replace the default integration" in normalized
         assert "specify integration switch claude" in normalized
         assert "To install 'claude' alongside" in normalized
@@ -358,6 +386,10 @@ class TestIntegrationInstall:
         # Shared infrastructure should be present
         assert (project / ".specify" / "scripts").is_dir()
         assert (project / ".specify" / "templates").is_dir()
+        script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+        script_content = script.read_text(encoding="utf-8")
+        assert "/speckit-specify" in script_content
+        assert "/speckit.specify" not in script_content
 
 
 # ── uninstall ────────────────────────────────────────────────────────
@@ -486,7 +518,9 @@ class TestIntegrationUninstall:
     def test_uninstall_default_refreshes_templates_for_fallback(self, tmp_path):
         project = _init_project(tmp_path, "gemini")
         template = project / ".specify" / "templates" / "plan-template.md"
+        script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
         assert "/speckit.plan" in template.read_text(encoding="utf-8")
+        assert "/speckit.plan" in script.read_text(encoding="utf-8")
 
         old_cwd = os.getcwd()
         try:
@@ -505,6 +539,7 @@ class TestIntegrationUninstall:
         data = json.loads((project / ".specify" / "integration.json").read_text(encoding="utf-8"))
         assert data["integration"] == "claude"
         assert "/speckit-plan" in template.read_text(encoding="utf-8")
+        assert "/speckit-plan" in script.read_text(encoding="utf-8")
 
     def test_uninstall_preserves_shared_infra(self, tmp_path):
         """Shared scripts and templates are not removed by integration uninstall."""
@@ -565,7 +600,9 @@ class TestIntegrationUse:
     def test_use_refreshes_shared_templates_between_command_styles(self, tmp_path):
         project = _init_project(tmp_path, "claude")
         template = project / ".specify" / "templates" / "plan-template.md"
+        script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
         assert "/speckit-plan" in template.read_text(encoding="utf-8")
+        assert "/speckit-plan" in script.read_text(encoding="utf-8")
 
         old_cwd = os.getcwd()
         try:
@@ -579,10 +616,14 @@ class TestIntegrationUse:
             use_gemini = runner.invoke(app, ["integration", "use", "gemini"], catch_exceptions=False)
             assert use_gemini.exit_code == 0, use_gemini.output
             assert "/speckit.plan" in template.read_text(encoding="utf-8")
+            assert "/speckit.plan" in script.read_text(encoding="utf-8")
+            assert "/speckit-plan" not in script.read_text(encoding="utf-8")
 
             use_claude = runner.invoke(app, ["integration", "use", "claude"], catch_exceptions=False)
             assert use_claude.exit_code == 0, use_claude.output
             assert "/speckit-plan" in template.read_text(encoding="utf-8")
+            assert "/speckit-plan" in script.read_text(encoding="utf-8")
+            assert "/speckit.plan" not in script.read_text(encoding="utf-8")
         finally:
             os.chdir(old_cwd)
 
@@ -602,6 +643,8 @@ class TestIntegrationUse:
 
             use_gemini = runner.invoke(app, ["integration", "use", "gemini"], catch_exceptions=False)
             assert use_gemini.exit_code == 0, use_gemini.output
+            normalized = " ".join(use_gemini.output.split())
+            assert "specify integration use gemini --force" in normalized
             assert template.read_text(encoding="utf-8") == "custom template with /speckit-plan\n"
 
             force_use = runner.invoke(app, [
@@ -616,8 +659,7 @@ class TestIntegrationUse:
         assert "/speckit.plan" in updated
         assert "custom template" not in updated
 
-    @pytest.mark.skipif(not hasattr(os, "symlink"), reason="symlinks are unavailable")
-    def test_use_does_not_persist_default_when_template_refresh_fails(self, tmp_path):
+    def test_use_does_not_persist_default_when_shared_infra_refresh_fails(self, tmp_path, monkeypatch):
         project = _init_project(tmp_path, "claude")
         int_json = project / ".specify" / "integration.json"
         init_options = project / ".specify" / "init-options.json"
@@ -633,12 +675,12 @@ class TestIntegrationUse:
 
             before_state = json.loads(int_json.read_text(encoding="utf-8"))
             before_options = json.loads(init_options.read_text(encoding="utf-8"))
+            import specify_cli
 
-            outside = tmp_path / "outside-template.md"
-            outside.write_text("# outside\n", encoding="utf-8")
-            template = project / ".specify" / "templates" / "plan-template.md"
-            template.unlink()
-            os.symlink(outside, template)
+            def fail_refresh(*args, **kwargs):
+                raise ValueError("refuse refresh")
+
+            monkeypatch.setattr(specify_cli, "_install_shared_infra", fail_refresh)
 
             result = runner.invoke(app, [
                 "integration", "use", "codex",
@@ -648,10 +690,9 @@ class TestIntegrationUse:
             os.chdir(old_cwd)
 
         assert result.exit_code != 0
-        assert "Failed to refresh shared templates" in result.output
+        assert "Failed to refresh shared infrastructure" in result.output
         assert json.loads(int_json.read_text(encoding="utf-8")) == before_state
         assert json.loads(init_options.read_text(encoding="utf-8")) == before_options
-        assert outside.read_text(encoding="utf-8") == "# outside\n"
 
 
 # ── switch ───────────────────────────────────────────────────────────
@@ -709,7 +750,9 @@ class TestIntegrationSwitch:
     def test_switch_same_force_refreshes_shared_templates(self, tmp_path):
         project = _init_project(tmp_path, "claude")
         template = project / ".specify" / "templates" / "plan-template.md"
+        script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
         template.write_text("# custom shared template\n", encoding="utf-8")
+        script.write_text("# custom shared script\n", encoding="utf-8")
 
         old_cwd = os.getcwd()
         try:
@@ -721,8 +764,10 @@ class TestIntegrationSwitch:
         finally:
             os.chdir(old_cwd)
         assert result.exit_code == 0, result.output
-        assert "managed shared templates refreshed" in result.output
+        assert "shared infrastructure refreshed" in result.output
+        assert "managed shared infrastructure refreshed" not in result.output
         assert "/speckit-plan" in template.read_text(encoding="utf-8")
+        assert "/speckit-plan" in script.read_text(encoding="utf-8")
 
     def test_switch_installed_target_rejects_integration_options(self, tmp_path):
         project = _init_project(tmp_path, "claude")
@@ -751,6 +796,8 @@ class TestIntegrationSwitch:
         project = _init_project(tmp_path, "claude")
         # Verify claude files exist (claude uses skills)
         assert (project / ".claude" / "skills" / "speckit-plan" / "SKILL.md").exists()
+        shared_script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+        assert "/speckit-specify" in shared_script.read_text(encoding="utf-8")
 
         old_cwd = os.getcwd()
         try:
@@ -769,6 +816,8 @@ class TestIntegrationSwitch:
 
         # New copilot files created
         assert (project / ".github" / "agents" / "speckit.plan.agent.md").exists()
+        assert "/speckit.specify" in shared_script.read_text(encoding="utf-8")
+        assert "/speckit-specify" not in shared_script.read_text(encoding="utf-8")
 
         # integration.json updated
         data = json.loads((project / ".specify" / "integration.json").read_text(encoding="utf-8"))
@@ -910,12 +959,13 @@ class TestIntegrationSwitch:
         assert "claude" not in git_meta["registered_commands"]
         assert "opencode" not in git_meta["registered_commands"]
 
-    def test_switch_preserves_shared_infra(self, tmp_path):
-        """Switching preserves shared scripts, templates, and memory."""
+    def test_switch_refreshes_managed_shared_script_refs(self, tmp_path):
+        """Switching refreshes managed shared scripts to the target command style."""
         project = _init_project(tmp_path, "claude")
         shared_script = project / ".specify" / "scripts" / "bash" / "common.sh"
         assert shared_script.exists()
         shared_content = shared_script.read_text(encoding="utf-8")
+        assert "/speckit-plan" in shared_content
 
         old_cwd = os.getcwd()
         try:
@@ -928,9 +978,10 @@ class TestIntegrationSwitch:
             os.chdir(old_cwd)
         assert result.exit_code == 0
 
-        # Shared infra untouched
         assert shared_script.exists()
-        assert shared_script.read_text(encoding="utf-8") == shared_content
+        updated = shared_script.read_text(encoding="utf-8")
+        assert "/speckit.plan" in updated
+        assert "/speckit-plan" not in updated
 
     def test_switch_refreshes_stale_managed_shared_infra(self, tmp_path):
         """Regression for #2293: stale managed shared scripts get refreshed on switch."""
@@ -938,7 +989,7 @@ class TestIntegrationSwitch:
 
         project = _init_project(tmp_path, "claude")
         shared_script = project / ".specify" / "scripts" / "bash" / "common.sh"
-        bundled_bytes = shared_script.read_bytes()
+        assert "/speckit-plan" in shared_script.read_text(encoding="utf-8")
 
         # Simulate a stale vendored script: write truncated content as bytes
         # (write_text would translate \n→\r\n on Windows and break the hash)
@@ -965,8 +1016,11 @@ class TestIntegrationSwitch:
             os.chdir(old_cwd)
         assert result.exit_code == 0
 
-        # Stale managed file should be replaced by the bundled version
-        assert shared_script.read_bytes() == bundled_bytes
+        # Stale managed file should be replaced by the target integration's rendered version.
+        updated = shared_script.read_text(encoding="utf-8")
+        assert "# stale vendored copy" not in updated
+        assert "/speckit.plan" in updated
+        assert "/speckit-plan" not in updated
 
     def test_switch_preserves_user_customized_shared_infra(self, tmp_path):
         """User customizations (hash divergence from manifest) survive switch without --refresh-shared-infra."""
@@ -996,10 +1050,11 @@ class TestIntegrationSwitch:
         """--refresh-shared-infra explicitly overwrites user customizations on switch."""
         project = _init_project(tmp_path, "claude")
         shared_script = project / ".specify" / "scripts" / "bash" / "common.sh"
-        bundled_bytes = shared_script.read_bytes()
+        assert "/speckit-plan" in shared_script.read_text(encoding="utf-8")
+        rendered_bytes = shared_script.read_bytes()
 
         # User customization (hash diverges from manifest)
-        custom_bytes = bundled_bytes + b"\n# user customization\n"
+        custom_bytes = rendered_bytes + b"\n# user customization\n"
         shared_script.write_bytes(custom_bytes)
 
         old_cwd = os.getcwd()
@@ -1013,8 +1068,11 @@ class TestIntegrationSwitch:
         finally:
             os.chdir(old_cwd)
         assert result.exit_code == 0
-        # Customization is overwritten with the bundled version
-        assert shared_script.read_bytes() == bundled_bytes
+        # Customization is overwritten with the target integration's rendered version.
+        updated = shared_script.read_text(encoding="utf-8")
+        assert "# user customization" not in updated
+        assert "/speckit.plan" in updated
+        assert "/speckit-plan" not in updated
 
     def test_switch_skips_symlinked_parent_directory(self, tmp_path):
         """Regression: if .specify/scripts/bash is a symlink, switch must not write through it.
@@ -1144,7 +1202,57 @@ class TestIntegrationUpgrade:
         assert "manifest" in result.output
         assert "unreadable" in result.output
 
-    def test_upgrade_does_not_persist_state_when_template_refresh_fails(self, tmp_path, monkeypatch):
+    def test_upgrade_refreshes_init_options_speckit_version(self, tmp_path, monkeypatch):
+        project = _init_project(tmp_path, "claude")
+        init_options = project / ".specify" / "init-options.json"
+        opts = json.loads(init_options.read_text(encoding="utf-8"))
+        opts["speckit_version"] = "0.6.1"
+        init_options.write_text(json.dumps(opts), encoding="utf-8")
+
+        import specify_cli
+
+        monkeypatch.setattr(specify_cli, "get_speckit_version", lambda: "0.8.11")
+
+        result = _run_in_project(project, [
+            "integration", "upgrade", "claude",
+            "--force",
+        ])
+
+        assert result.exit_code == 0, result.output
+        updated = json.loads(init_options.read_text(encoding="utf-8"))
+        assert updated["speckit_version"] == "0.8.11"
+
+    def test_upgrade_non_default_refreshes_init_options_version_only(self, tmp_path, monkeypatch):
+        project = _init_project(tmp_path, "gemini")
+        install = _run_in_project(project, [
+            "integration", "install", "claude",
+            "--script", "sh",
+        ])
+        assert install.exit_code == 0, install.output
+
+        init_options = project / ".specify" / "init-options.json"
+        opts = json.loads(init_options.read_text(encoding="utf-8"))
+        opts["speckit_version"] = "0.6.1"
+        init_options.write_text(json.dumps(opts), encoding="utf-8")
+
+        import specify_cli
+
+        monkeypatch.setattr(specify_cli, "get_speckit_version", lambda: "0.8.11")
+
+        result = _run_in_project(project, [
+            "integration", "upgrade", "claude",
+            "--script", "sh",
+            "--force",
+        ])
+
+        assert result.exit_code == 0, result.output
+        updated = json.loads(init_options.read_text(encoding="utf-8"))
+        assert updated["speckit_version"] == "0.8.11"
+        assert updated["integration"] == "gemini"
+        assert updated["ai"] == "gemini"
+        assert updated["context_file"] == "GEMINI.md"
+
+    def test_upgrade_does_not_persist_state_when_shared_infra_refresh_fails(self, tmp_path, monkeypatch):
         project = _init_project(tmp_path, "claude")
         int_json = project / ".specify" / "integration.json"
         init_options = project / ".specify" / "init-options.json"
@@ -1156,10 +1264,16 @@ class TestIntegrationUpgrade:
 
         import specify_cli
 
-        def fail_refresh(*args, **kwargs):
-            raise ValueError("refuse refresh")
+        real_install_shared_infra = specify_cli._install_shared_infra
+        calls = {"count": 0}
 
-        monkeypatch.setattr(specify_cli, "_refresh_shared_templates", fail_refresh)
+        def fail_refresh(*args, **kwargs):
+            calls["count"] += 1
+            if calls["count"] == 2:
+                raise ValueError("refuse refresh")
+            return real_install_shared_infra(*args, **kwargs)
+
+        monkeypatch.setattr(specify_cli, "_install_shared_infra", fail_refresh)
 
         result = _run_in_project(project, [
             "integration", "upgrade", "claude",
@@ -1167,15 +1281,40 @@ class TestIntegrationUpgrade:
         ])
 
         assert result.exit_code != 0
-        assert "Failed to refresh shared templates" in result.output
+        assert "Failed to refresh shared infrastructure" in result.output
         assert json.loads(int_json.read_text(encoding="utf-8")) == before_state
         assert json.loads(init_options.read_text(encoding="utf-8")) == before_options
         assert manifest_path.read_text(encoding="utf-8") == before_manifest
 
+    def test_upgrade_default_refreshes_shared_script_refs_for_option_separator_change(self, tmp_path):
+        project = _init_project(tmp_path, "copilot")
+        template = project / ".specify" / "templates" / "plan-template.md"
+        managed_script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+        customized_script = project / ".specify" / "scripts" / "bash" / "setup-tasks.sh"
+
+        assert "/speckit.plan" in template.read_text(encoding="utf-8")
+        assert "/speckit.specify" in managed_script.read_text(encoding="utf-8")
+        customized_before = customized_script.read_text(encoding="utf-8") + "\n# user customization\n"
+        customized_script.write_text(customized_before, encoding="utf-8")
+
+        result = _run_in_project(project, [
+            "integration", "upgrade", "copilot",
+            "--integration-options", "--skills",
+        ])
+
+        assert result.exit_code == 0, result.output
+        assert "/speckit-plan" in template.read_text(encoding="utf-8")
+        managed_content = managed_script.read_text(encoding="utf-8")
+        assert "/speckit-specify" in managed_content
+        assert "/speckit.specify" not in managed_content
+        assert customized_script.read_text(encoding="utf-8") == customized_before
+
     def test_upgrade_non_default_keeps_default_template_invocations(self, tmp_path):
         project = _init_project(tmp_path, "gemini")
         template = project / ".specify" / "templates" / "plan-template.md"
+        script = project / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
         assert "/speckit.plan" in template.read_text(encoding="utf-8")
+        assert "/speckit.plan" in script.read_text(encoding="utf-8")
 
         old_cwd = os.getcwd()
         try:
@@ -1198,6 +1337,8 @@ class TestIntegrationUpgrade:
         data = json.loads((project / ".specify" / "integration.json").read_text(encoding="utf-8"))
         assert data["integration"] == "gemini"
         assert "/speckit.plan" in template.read_text(encoding="utf-8")
+        assert "/speckit.plan" in script.read_text(encoding="utf-8")
+        assert "/speckit-plan" not in script.read_text(encoding="utf-8")
 
     def test_upgrade_migrates_opencode_legacy_dir(self, tmp_path):
         """Upgrade moves OpenCode commands from .opencode/command/ to .opencode/commands/."""

tests/test_check_prerequisites_paths_only.py

@@ -0,0 +1,205 @@
+"""Tests for check-prerequisites --paths-only skipping branch validation (#2653)."""
+
+import json
+import os
+import shutil
+import subprocess
+from pathlib import Path
+
+import pytest
+
+from tests.conftest import requires_bash
+
+PROJECT_ROOT = Path(__file__).resolve().parent.parent
+COMMON_SH = PROJECT_ROOT / "scripts" / "bash" / "common.sh"
+CHECK_PREREQS_SH = PROJECT_ROOT / "scripts" / "bash" / "check-prerequisites.sh"
+COMMON_PS = PROJECT_ROOT / "scripts" / "powershell" / "common.ps1"
+CHECK_PREREQS_PS = PROJECT_ROOT / "scripts" / "powershell" / "check-prerequisites.ps1"
+
+HAS_PWSH = shutil.which("pwsh") is not None
+_POWERSHELL = shutil.which("powershell.exe") or shutil.which("powershell")
+
+
+def _install_bash_scripts(repo: Path) -> None:
+    d = repo / ".specify" / "scripts" / "bash"
+    d.mkdir(parents=True, exist_ok=True)
+    shutil.copy(COMMON_SH, d / "common.sh")
+    shutil.copy(CHECK_PREREQS_SH, d / "check-prerequisites.sh")
+
+
+def _install_ps_scripts(repo: Path) -> None:
+    d = repo / ".specify" / "scripts" / "powershell"
+    d.mkdir(parents=True, exist_ok=True)
+    shutil.copy(COMMON_PS, d / "common.ps1")
+    shutil.copy(CHECK_PREREQS_PS, d / "check-prerequisites.ps1")
+
+
+def _clean_env() -> dict[str, str]:
+    env = os.environ.copy()
+    for key in list(env):
+        if key.startswith("SPECIFY_"):
+            env.pop(key)
+    return env
+
+
+def _git_init(repo: Path) -> None:
+    subprocess.run(["git", "init", "-q"], cwd=repo, check=True)
+    subprocess.run(
+        ["git", "config", "user.email", "test@example.com"], cwd=repo, check=True
+    )
+    subprocess.run(["git", "config", "user.name", "Test User"], cwd=repo, check=True)
+    subprocess.run(
+        ["git", "commit", "--allow-empty", "-m", "init", "-q"], cwd=repo, check=True
+    )
+
+
+@pytest.fixture
+def prereq_repo(tmp_path: Path) -> Path:
+    repo = tmp_path / "proj"
+    repo.mkdir()
+    _git_init(repo)
+    (repo / ".specify").mkdir()
+    _install_bash_scripts(repo)
+    _install_ps_scripts(repo)
+    return repo
+
+
+# ── Bash tests ────────────────────────────────────────────────────────────
+
+
+@requires_bash
+def test_paths_only_succeeds_on_non_spec_branch(prereq_repo: Path) -> None:
+    """--paths-only must return paths without branch validation (main branch)."""
+    script = prereq_repo / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json", "--paths-only"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    assert "REPO_ROOT" in data
+    assert "BRANCH" in data
+    assert "FEATURE_DIR" in data
+
+
+@requires_bash
+def test_paths_only_succeeds_on_spec_branch(prereq_repo: Path) -> None:
+    """--paths-only must also work on a properly named spec branch."""
+    subprocess.run(
+        ["git", "checkout", "-q", "-b", "001-my-feature"],
+        cwd=prereq_repo,
+        check=True,
+    )
+    script = prereq_repo / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json", "--paths-only"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    assert "FEATURE_DIR" in data
+    assert "001-my-feature" in data.get("BRANCH", "")
+
+
+@requires_bash
+def test_paths_only_text_mode_on_non_spec_branch(prereq_repo: Path) -> None:
+    """--paths-only without --json must return text paths on a non-spec branch."""
+    script = prereq_repo / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--paths-only"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    assert "REPO_ROOT:" in result.stdout
+    assert "FEATURE_DIR:" in result.stdout
+
+
+@requires_bash
+def test_normal_mode_still_validates_branch(prereq_repo: Path) -> None:
+    """Without --paths-only, branch validation must still fail on main."""
+    script = prereq_repo / ".specify" / "scripts" / "bash" / "check-prerequisites.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode != 0
+    assert "Not on a feature branch" in result.stderr
+
+
+# ── PowerShell tests ──────────────────────────────────────────────────────
+
+
+@pytest.mark.skipif(not (HAS_PWSH or _POWERSHELL), reason="no PowerShell available")
+def test_ps_paths_only_succeeds_on_non_spec_branch(prereq_repo: Path) -> None:
+    """-PathsOnly must return paths without branch validation (main branch)."""
+    script = prereq_repo / ".specify" / "scripts" / "powershell" / "check-prerequisites.ps1"
+    exe = "pwsh" if HAS_PWSH else _POWERSHELL
+    result = subprocess.run(
+        [exe, "-NoProfile", "-File", str(script), "-Json", "-PathsOnly"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    assert "REPO_ROOT" in data
+    assert "BRANCH" in data
+    assert "FEATURE_DIR" in data
+
+
+@pytest.mark.skipif(not (HAS_PWSH or _POWERSHELL), reason="no PowerShell available")
+def test_ps_paths_only_succeeds_on_spec_branch(prereq_repo: Path) -> None:
+    """-PathsOnly must also work on a properly named spec branch."""
+    subprocess.run(
+        ["git", "checkout", "-q", "-b", "001-my-feature"],
+        cwd=prereq_repo,
+        check=True,
+    )
+    script = prereq_repo / ".specify" / "scripts" / "powershell" / "check-prerequisites.ps1"
+    exe = "pwsh" if HAS_PWSH else _POWERSHELL
+    result = subprocess.run(
+        [exe, "-NoProfile", "-File", str(script), "-Json", "-PathsOnly"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    assert "FEATURE_DIR" in data
+
+
+@pytest.mark.skipif(not (HAS_PWSH or _POWERSHELL), reason="no PowerShell available")
+def test_ps_normal_mode_still_validates_branch(prereq_repo: Path) -> None:
+    """Without -PathsOnly, branch validation must still fail on main."""
+    script = prereq_repo / ".specify" / "scripts" / "powershell" / "check-prerequisites.ps1"
+    exe = "pwsh" if HAS_PWSH else _POWERSHELL
+    result = subprocess.run(
+        [exe, "-NoProfile", "-File", str(script), "-Json"],
+        cwd=prereq_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode != 0
+    assert "Not on a feature branch" in result.stderr

tests/test_check_tool.py

@@ -7,7 +7,13 @@ Covers issue https://github.com/github/spec-kit/issues/550:
 
 from unittest.mock import patch, MagicMock
 
-from specify_cli import check_tool
+from typer.testing import CliRunner
+
+from specify_cli import app, check_tool
+from tests.conftest import strip_ansi
+
+
+runner = CliRunner()
 
 
 class TestCheckToolClaude:
@@ -103,4 +109,32 @@ class TestCheckToolOther:
             return "/usr/bin/kiro" if name == "kiro" else None
 
         with patch("shutil.which", side_effect=fake_which):
-            assert check_tool("kiro-cli") is True
+            assert check_tool("kiro-cli") is True
+
+
+class TestCheckTip:
+    """`specify check` should point users to the existing version check."""
+
+    def test_check_shows_self_check_tip(self):
+        with patch("specify_cli.check_tool", return_value=True):
+            result = runner.invoke(app, ["check"])
+
+        output = strip_ansi(result.output)
+        assert result.exit_code == 0
+        assert (
+            "Tip: Run 'specify self check' to verify you have the latest CLI version"
+            in output
+        )
+
+    def test_check_tip_does_not_fetch_latest_release(self):
+        with (
+            patch("specify_cli.check_tool", return_value=True),
+            patch(
+                "specify_cli._version._fetch_latest_release_tag",
+                side_effect=AssertionError("latest release lookup should not run"),
+            ) as fetch_latest,
+        ):
+            result = runner.invoke(app, ["check"])
+
+        assert result.exit_code == 0
+        fetch_latest.assert_not_called()

tests/test_commands_package.py

@@ -0,0 +1,48 @@
+"""Tests for the commands/ package structure."""
+import importlib
+
+
+def test_commands_package_importable():
+    mod = importlib.import_module("specify_cli.commands")
+    assert mod is not None
+
+
+def test_commands_init_importable():
+    mod = importlib.import_module("specify_cli.commands.init")
+    assert hasattr(mod, "register")
+    assert callable(mod.register)
+
+
+def test_commands_stubs_importable():
+    for name in ("integration", "preset", "extension", "workflow"):
+        mod = importlib.import_module(f"specify_cli.commands.{name}")
+        assert mod is not None
+
+
+def test_agent_config_importable():
+    from specify_cli._agent_config import (
+        AGENT_CONFIG,
+        AI_ASSISTANT_ALIASES,
+        AI_ASSISTANT_HELP,
+        DEFAULT_INIT_INTEGRATION,
+        SCRIPT_TYPE_CHOICES,
+    )
+    assert isinstance(AGENT_CONFIG, dict)
+    assert isinstance(AI_ASSISTANT_ALIASES, dict)
+    assert isinstance(AI_ASSISTANT_HELP, str)
+    assert DEFAULT_INIT_INTEGRATION == "copilot"
+    assert "sh" in SCRIPT_TYPE_CHOICES
+
+
+def test_agent_config_re_exported_from_init():
+    from specify_cli import AGENT_CONFIG, AI_ASSISTANT_ALIASES, AI_ASSISTANT_HELP, SCRIPT_TYPE_CHOICES
+    assert isinstance(AGENT_CONFIG, dict)
+    assert "sh" in SCRIPT_TYPE_CHOICES
+
+
+def test_init_command_registered():
+    from specify_cli import app
+    callback_names = [
+        cmd.callback.__name__ for cmd in app.registered_commands if cmd.callback
+    ]
+    assert "init" in callback_names

tests/test_extension_skills.py

@@ -11,6 +11,7 @@ Tests cover:
 """
 
 import json
+import os
 import pytest
 import tempfile
 import shutil
@@ -116,6 +117,18 @@ def _create_extension_dir(temp_dir: Path, ext_id: str = "test-ext") -> Path:
     return ext_dir
 
 
+def _can_create_symlink(temp_dir: Path) -> bool:
+    """Return True when the current platform/user can create file symlinks."""
+    target = temp_dir / "symlink-target.txt"
+    link = temp_dir / "symlink-link.txt"
+    target.write_text("ok", encoding="utf-8")
+    try:
+        os.symlink(target, link)
+    except OSError:
+        return False
+    return link.is_symlink()
+
+
 # ===== Fixtures =====
 
 @pytest.fixture
@@ -173,24 +186,32 @@ class TestExtensionManagerGetSkillsDir:
         assert result == skills_dir
 
     def test_returns_none_when_no_ai_skills(self, no_skills_project):
-        """Should return None when ai_skills is false."""
+        """Should return None when ai_skills is false and not create the dir."""
         manager = ExtensionManager(no_skills_project)
         result = manager._get_skills_dir()
         assert result is None
+        # Ensure the directory was NOT created on disk
+        from specify_cli import _get_skills_dir as resolve_skills_dir
+        skills_path = resolve_skills_dir(no_skills_project, "claude")
+        assert not skills_path.exists()
 
     def test_returns_none_when_no_init_options(self, project_dir):
-        """Should return None when init-options.json is missing."""
+        """Should return None when init-options.json is missing and not create any dir."""
         manager = ExtensionManager(project_dir)
         result = manager._get_skills_dir()
         assert result is None
+        # No agent skills directory should have been created
+        assert not (project_dir / ".claude" / "skills").exists()
+        assert not (project_dir / ".agents" / "skills").exists()
 
-    def test_returns_none_when_skills_dir_missing(self, project_dir):
-        """Should return None when skills dir doesn't exist on disk."""
+    def test_creates_skills_dir_on_demand(self, project_dir):
+        """Should create skills dir when ai_skills is enabled but dir is missing."""
         _create_init_options(project_dir, ai="claude", ai_skills=True)
-        # Don't create the skills directory
+        # Don't create the skills directory — _get_skills_dir should do it
         manager = ExtensionManager(project_dir)
         result = manager._get_skills_dir()
-        assert result is None
+        assert result is not None
+        assert result.is_dir()
 
     def test_returns_kimi_skills_dir_when_ai_skills_disabled(self, project_dir):
         """Kimi should still use its native skills dir when ai_skills is false."""
@@ -316,6 +337,149 @@ class TestExtensionSkillRegistration:
         # The pre-existing one should NOT be in registered_skills (it was skipped)
         assert "speckit-test-ext-hello" not in metadata["registered_skills"]
 
+    def test_dev_skill_symlink_refreshes_existing_cache(
+        self, skills_project, extension_dir, temp_dir
+    ):
+        """Dev-mode skill symlinks should refresh rendered cache content."""
+        if not _can_create_symlink(temp_dir):
+            pytest.skip("Current platform/user cannot create symlinks")
+
+        project_dir, skills_dir = skills_project
+        manager = ExtensionManager(project_dir)
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+
+        manager._register_extension_skills(
+            manifest,
+            extension_dir,
+            link_outputs=True,
+        )
+
+        skill_file = skills_dir / "speckit-test-ext-hello" / "SKILL.md"
+        assert skill_file.is_symlink()
+        assert "Run this to say hello." in skill_file.read_text(encoding="utf-8")
+
+        (extension_dir / "commands" / "hello.md").write_text(
+            "---\n"
+            "description: \"Updated test hello command\"\n"
+            "---\n"
+            "\n"
+            "# Hello Command\n"
+            "\n"
+            "Run this updated hello.\n"
+        )
+
+        written = manager._register_extension_skills(
+            manifest,
+            extension_dir,
+            link_outputs=True,
+        )
+
+        assert "speckit-test-ext-hello" in written
+        assert "Run this updated hello." in skill_file.read_text(encoding="utf-8")
+
+    def test_dev_skill_registration_falls_back_to_copy_when_symlink_fails(
+        self, skills_project, extension_dir, monkeypatch
+    ):
+        """Dev-mode skill registration works when Windows cannot create symlinks."""
+        project_dir, skills_dir = skills_project
+        manager = ExtensionManager(project_dir)
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+
+        def raise_windows_symlink_error(target, link):
+            raise OSError("A required privilege is not held by the client")
+
+        monkeypatch.setattr(
+            "specify_cli.extensions.os.symlink", raise_windows_symlink_error
+        )
+
+        written = manager._register_extension_skills(
+            manifest,
+            extension_dir,
+            link_outputs=True,
+        )
+
+        skill_file = skills_dir / "speckit-test-ext-hello" / "SKILL.md"
+        assert "speckit-test-ext-hello" in written
+        assert skill_file.exists()
+        assert not skill_file.is_symlink()
+        assert "Run this to say hello." in skill_file.read_text(encoding="utf-8")
+        assert (
+            extension_dir
+            / ".specify-dev"
+            / "extension-skills"
+            / "speckit-test-ext-hello"
+            / "SKILL.md"
+        ).exists()
+
+    def test_dev_skill_registration_falls_back_to_copy_when_relpath_fails(
+        self, skills_project, extension_dir, monkeypatch
+    ):
+        """Dev-mode skill registration stays functional across Windows drive roots."""
+        project_dir, skills_dir = skills_project
+        manager = ExtensionManager(project_dir)
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+
+        def raise_relpath_error(path, start=None):
+            raise ValueError("path is on mount 'D:', start on mount 'C:'")
+
+        monkeypatch.setattr(
+            "specify_cli.extensions.os.path.relpath", raise_relpath_error
+        )
+
+        written = manager._register_extension_skills(
+            manifest,
+            extension_dir,
+            link_outputs=True,
+        )
+
+        skill_file = skills_dir / "speckit-test-ext-hello" / "SKILL.md"
+        assert "speckit-test-ext-hello" in written
+        assert skill_file.exists()
+        assert not skill_file.is_symlink()
+        assert "Run this to say hello." in skill_file.read_text(encoding="utf-8")
+        assert (
+            extension_dir
+            / ".specify-dev"
+            / "extension-skills"
+            / "speckit-test-ext-hello"
+            / "SKILL.md"
+        ).exists()
+
+    def test_dev_skill_registration_falls_back_to_copy_when_cache_write_fails(
+        self, skills_project, extension_dir, monkeypatch
+    ):
+        """Dev-mode skill registration stays functional when the dev cache is unwritable."""
+        project_dir, skills_dir = skills_project
+        manager = ExtensionManager(project_dir)
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+        original_write_text = Path.write_text
+
+        def raise_cache_write_error(path, *args, **kwargs):
+            if ".specify-dev" in path.parts:
+                raise OSError("cache is not writable")
+            return original_write_text(path, *args, **kwargs)
+
+        monkeypatch.setattr(Path, "write_text", raise_cache_write_error)
+
+        written = manager._register_extension_skills(
+            manifest,
+            extension_dir,
+            link_outputs=True,
+        )
+
+        skill_file = skills_dir / "speckit-test-ext-hello" / "SKILL.md"
+        assert "speckit-test-ext-hello" in written
+        assert skill_file.exists()
+        assert not skill_file.is_symlink()
+        assert "Run this to say hello." in skill_file.read_text(encoding="utf-8")
+        assert not (
+            extension_dir
+            / ".specify-dev"
+            / "extension-skills"
+            / "speckit-test-ext-hello"
+            / "SKILL.md"
+        ).exists()
+
     def test_registered_skills_in_registry(self, skills_project, extension_dir):
         """Registry should contain registered_skills list."""
         project_dir, skills_dir = skills_project
@@ -460,6 +624,38 @@ class TestExtensionSkillRegistration:
         assert "speckit-missing-cmd-ext-exists" in metadata["registered_skills"]
         assert "speckit-missing-cmd-ext-ghost" not in metadata["registered_skills"]
 
+    @pytest.mark.parametrize("ai", ["claude", "codex"])
+    def test_skills_registered_when_dir_missing(self, project_dir, temp_dir, ai):
+        """Extension add should create skills dir on demand and register skills.
+
+        Regression test for https://github.com/github/spec-kit/issues/2682:
+        when an extension is installed before the agent skills directory exists,
+        skills must still be materialized (the directory is created on demand).
+        """
+        _create_init_options(project_dir, ai=ai, ai_skills=True)
+        # Deliberately do NOT create the skills directory
+        ext_dir = _create_extension_dir(temp_dir, ext_id="early-ext")
+
+        manager = ExtensionManager(project_dir)
+        manifest = manager.install_from_directory(
+            ext_dir, "0.1.0", register_commands=False
+        )
+
+        # Skills dir should have been created automatically
+        from specify_cli import _get_skills_dir as resolve_skills_dir
+        skills_dir = resolve_skills_dir(project_dir, ai)
+        assert skills_dir.is_dir()
+
+        # SKILL.md files should exist
+        assert (skills_dir / "speckit-early-ext-hello" / "SKILL.md").exists()
+        assert (skills_dir / "speckit-early-ext-world" / "SKILL.md").exists()
+
+        # Registry should record them
+        metadata = manager.registry.get(manifest.id)
+        assert len(metadata["registered_skills"]) == 2
+        assert "speckit-early-ext-hello" in metadata["registered_skills"]
+        assert "speckit-early-ext-world" in metadata["registered_skills"]
+
 
 # ===== Extension Skill Unregistration Tests =====
 

tests/test_extensions.py

@@ -11,6 +11,7 @@ Tests cover:
 
 import pytest
 import json
+import os
 import platform
 import tempfile
 import shutil
@@ -36,6 +37,18 @@ from specify_cli.extensions import (
 )
 
 
+def can_create_symlink(tmp_path: Path) -> bool:
+    """Return True when the current platform/user can create file symlinks."""
+    target = tmp_path / "symlink-target.txt"
+    link = tmp_path / "symlink-link.txt"
+    target.write_text("ok", encoding="utf-8")
+    try:
+        os.symlink(target, link)
+    except OSError:
+        return False
+    return link.is_symlink()
+
+
 # ===== Fixtures =====
 
 @pytest.fixture
@@ -1722,6 +1735,168 @@ Run {SCRIPT}
         assert "description: Test hello command" in content
         assert "test-ext" in content
 
+    def test_dev_register_commands_symlinks_rendered_copilot_agent(
+        self, extension_dir, project_dir, temp_dir
+    ):
+        """Dev-mode registration should symlink agent files to rendered outputs."""
+        if not can_create_symlink(temp_dir):
+            pytest.skip("Current platform/user cannot create symlinks")
+
+        agents_dir = project_dir / ".github" / "agents"
+        agents_dir.mkdir(parents=True)
+
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+        registrar = CommandRegistrar()
+        registered = registrar.register_commands_for_agent(
+            "copilot",
+            manifest,
+            extension_dir,
+            project_dir,
+            link_outputs=True,
+        )
+
+        assert registered == ["speckit.test-ext.hello"]
+
+        cmd_file = agents_dir / "speckit.test-ext.hello.agent.md"
+        assert cmd_file.is_symlink()
+
+        target = cmd_file.resolve()
+        assert ".specify-dev" in target.parts
+        assert target.is_file()
+        assert "Extension: test-ext" in cmd_file.read_text(encoding="utf-8")
+
+    def test_dev_register_commands_falls_back_to_copy_when_symlink_fails(
+        self, extension_dir, project_dir, monkeypatch
+    ):
+        """Dev-mode registration stays functional when symlinks are unavailable."""
+        agents_dir = project_dir / ".github" / "agents"
+        agents_dir.mkdir(parents=True)
+
+        def raise_symlink_error(target, link):
+            raise OSError("symlink unavailable")
+
+        monkeypatch.setattr("specify_cli.agents.os.symlink", raise_symlink_error)
+
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+        registrar = CommandRegistrar()
+        registrar.register_commands_for_agent(
+            "copilot",
+            manifest,
+            extension_dir,
+            project_dir,
+            link_outputs=True,
+        )
+
+        cmd_file = agents_dir / "speckit.test-ext.hello.agent.md"
+        assert cmd_file.exists()
+        assert not cmd_file.is_symlink()
+        assert "Extension: test-ext" in cmd_file.read_text(encoding="utf-8")
+        assert (
+            extension_dir
+            / ".specify-dev"
+            / "agent-commands"
+            / "copilot"
+            / "speckit.test-ext.hello.agent.md"
+        ).exists()
+
+    def test_dev_register_commands_falls_back_to_copy_when_relpath_fails(
+        self, extension_dir, project_dir, monkeypatch
+    ):
+        """Dev-mode registration stays functional across Windows drive roots."""
+        agents_dir = project_dir / ".github" / "agents"
+        agents_dir.mkdir(parents=True)
+
+        def raise_relpath_error(path, start=None):
+            raise ValueError("path is on mount 'D:', start on mount 'C:'")
+
+        monkeypatch.setattr("specify_cli.agents.os.path.relpath", raise_relpath_error)
+
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+        registrar = CommandRegistrar()
+        registrar.register_commands_for_agent(
+            "copilot",
+            manifest,
+            extension_dir,
+            project_dir,
+            link_outputs=True,
+        )
+
+        cmd_file = agents_dir / "speckit.test-ext.hello.agent.md"
+        assert cmd_file.exists()
+        assert not cmd_file.is_symlink()
+        assert "Extension: test-ext" in cmd_file.read_text(encoding="utf-8")
+        assert (
+            extension_dir
+            / ".specify-dev"
+            / "agent-commands"
+            / "copilot"
+            / "speckit.test-ext.hello.agent.md"
+        ).exists()
+
+    def test_dev_register_commands_falls_back_to_copy_when_cache_write_fails(
+        self, extension_dir, project_dir, monkeypatch
+    ):
+        """Dev-mode registration stays functional when the dev cache is unwritable."""
+        agents_dir = project_dir / ".github" / "agents"
+        agents_dir.mkdir(parents=True)
+        original_write_text = Path.write_text
+
+        def raise_cache_write_error(path, *args, **kwargs):
+            if ".specify-dev" in path.parts:
+                raise OSError("cache is not writable")
+            return original_write_text(path, *args, **kwargs)
+
+        monkeypatch.setattr(Path, "write_text", raise_cache_write_error)
+
+        manifest = ExtensionManifest(extension_dir / "extension.yml")
+        registrar = CommandRegistrar()
+        registrar.register_commands_for_agent(
+            "copilot",
+            manifest,
+            extension_dir,
+            project_dir,
+            link_outputs=True,
+        )
+
+        cmd_file = agents_dir / "speckit.test-ext.hello.agent.md"
+        assert cmd_file.exists()
+        assert not cmd_file.is_symlink()
+        assert "Extension: test-ext" in cmd_file.read_text(encoding="utf-8")
+        assert not (
+            extension_dir
+            / ".specify-dev"
+            / "agent-commands"
+            / "copilot"
+            / "speckit.test-ext.hello.agent.md"
+        ).exists()
+
+    def test_dev_register_commands_rejects_cache_path_traversal(self, temp_dir):
+        """Dev-mode cache writes must stay inside the agent cache root."""
+        from specify_cli.agents import CommandRegistrar as AgentCommandRegistrar
+
+        source_dir = temp_dir / "extension"
+        source_dir.mkdir()
+        commands_dir = temp_dir / "commands"
+        commands_dir.mkdir()
+
+        with pytest.raises(ValueError, match="escapes directory"):
+            AgentCommandRegistrar._write_registered_output(
+                commands_dir / "safe.md",
+                "content",
+                source_dir,
+                "copilot",
+                "../escaped",
+                ".md",
+                True,
+            )
+
+        assert not (
+            source_dir
+            / ".specify-dev"
+            / "agent-commands"
+            / "escaped.md"
+        ).exists()
+
     def test_copilot_companion_prompt_created(self, extension_dir, project_dir):
         """Test that companion .prompt.md files are created in .github/prompts/."""
         agents_dir = project_dir / ".github" / "agents"
@@ -3458,6 +3633,86 @@ class TestExtensionIgnore:
 class TestExtensionAddCLI:
     """CLI integration tests for extension add command."""
 
+    def test_add_dev_links_copilot_agent_when_supported(
+        self, extension_dir, project_dir, temp_dir
+    ):
+        """extension add --dev should link generated agent files when possible."""
+        from typer.testing import CliRunner
+        from unittest.mock import patch
+        from specify_cli import app
+
+        (project_dir / ".github" / "agents").mkdir(parents=True)
+
+        runner = CliRunner()
+        with patch.object(Path, "cwd", return_value=project_dir):
+            result = runner.invoke(
+                app,
+                ["extension", "add", str(extension_dir), "--dev"],
+                catch_exceptions=True,
+            )
+
+        assert result.exit_code == 0, result.output
+
+        agent_file = (
+            project_dir
+            / ".github"
+            / "agents"
+            / "speckit.test-ext.hello.agent.md"
+        )
+        assert agent_file.exists()
+        if can_create_symlink(temp_dir):
+            assert agent_file.is_symlink()
+            assert ".specify-dev" in agent_file.resolve().parts
+        else:
+            assert not agent_file.is_symlink()
+
+    def test_add_dev_falls_back_to_copy_when_windows_symlinks_unavailable(
+        self, extension_dir, project_dir, monkeypatch
+    ):
+        """extension add --dev should work when Windows cannot create symlinks."""
+        from typer.testing import CliRunner
+        from unittest.mock import patch
+        from specify_cli import app
+
+        (project_dir / ".github" / "agents").mkdir(parents=True)
+
+        def raise_windows_symlink_error(target, link):
+            raise OSError("A required privilege is not held by the client")
+
+        monkeypatch.setattr(
+            "specify_cli.agents.os.symlink", raise_windows_symlink_error
+        )
+
+        runner = CliRunner()
+        with patch.object(Path, "cwd", return_value=project_dir):
+            result = runner.invoke(
+                app,
+                ["extension", "add", str(extension_dir), "--dev"],
+                catch_exceptions=True,
+            )
+
+        assert result.exit_code == 0, result.output
+
+        agent_file = (
+            project_dir
+            / ".github"
+            / "agents"
+            / "speckit.test-ext.hello.agent.md"
+        )
+        assert agent_file.exists()
+        assert not agent_file.is_symlink()
+        assert "Extension: test-ext" in agent_file.read_text(encoding="utf-8")
+        assert (
+            project_dir
+            / ".specify"
+            / "extensions"
+            / "test-ext"
+            / ".specify-dev"
+            / "agent-commands"
+            / "copilot"
+            / "speckit.test-ext.hello.agent.md"
+        ).exists()
+
     def test_add_by_display_name_uses_resolved_id_for_download(self, tmp_path):
         """extension add by display name should use resolved ID for download_extension()."""
         from typer.testing import CliRunner
@@ -3744,13 +3999,20 @@ class TestExtensionUpdateCLI:
         ).read_text()
         assert restored_config_content == original_config_content
 
-    def test_update_failure_rolls_back_registry_hooks_and_commands(self, tmp_path):
+    def test_update_failure_rolls_back_registry_hooks_and_commands(self, tmp_path, monkeypatch):
         """Failed update should restore original registry, hooks, and command files."""
         from typer.testing import CliRunner
         from unittest.mock import patch
         from specify_cli import app
         import yaml
 
+        # Isolate home directory so Hermes' global ~/.hermes/skills/ doesn't
+        # interfere — without a real skills dir, Hermes is skipped during
+        # command registration, keeping the test focused on Claude/Codex/etc.
+        fake_home = tmp_path / "home"
+        fake_home.mkdir()
+        monkeypatch.setattr(Path, "home", lambda: fake_home)
+
         runner = CliRunner()
         project_dir = tmp_path / "project"
         project_dir.mkdir()
@@ -3772,7 +4034,9 @@ class TestExtensionUpdateCLI:
             if agent_name not in agent_registrar.AGENT_CONFIGS:
                 continue
             agent_cfg = agent_registrar.AGENT_CONFIGS[agent_name]
-            commands_dir = project_dir / agent_cfg["dir"]
+            commands_dir = AgentRegistrar._resolve_agent_dir(
+                agent_name, agent_cfg, project_dir
+            )
             for cmd_name in cmd_names:
                 output_name = AgentRegistrar._compute_output_name(agent_name, cmd_name, agent_cfg)
                 cmd_path = commands_dir / f"{output_name}{agent_cfg['extension']}"

tests/test_presets.py

@@ -1830,6 +1830,31 @@ class TestPresetCatalogMultiCatalog:
         with pytest.raises(PresetValidationError, match="Invalid priority"):
             catalog._load_catalog_config(config_path)
 
+    def test_load_catalog_config_rejects_boolean_priority(self, project_dir):
+        """A YAML ``priority: true`` is a typo, not a request for priority 1.
+
+        ``bool`` is a subclass of ``int`` in Python, so ``int(True)`` silently
+        returns ``1``. Without an explicit guard a malformed config like
+        ``priority: yes`` would be accepted as a valid priority of 1 and
+        silently change catalog ordering. The sibling integration-catalog
+        reader rejects this case (see ``catalogs.py``); the preset catalog
+        reader must stay consistent.
+        """
+        config_path = project_dir / ".specify" / "preset-catalogs.yml"
+        config_path.write_text(yaml.dump({
+            "catalogs": [
+                {
+                    "name": "bool-priority",
+                    "url": "https://example.com/catalog.json",
+                    "priority": True,
+                }
+            ]
+        }))
+
+        catalog = PresetCatalog(project_dir)
+        with pytest.raises(PresetValidationError, match="Invalid priority|expected integer"):
+            catalog._load_catalog_config(config_path)
+
     def test_load_catalog_config_install_allowed_string(self, project_dir):
         """Test that install_allowed accepts string values."""
         config_path = project_dir / ".specify" / "preset-catalogs.yml"
@@ -2321,6 +2346,154 @@ class TestPresetSkills:
         metadata = manager.registry.get("self-test")
         assert "speckit-specify" in metadata.get("registered_skills", [])
 
+    def test_register_skills_resolves_command_refs(self, project_dir, temp_dir):
+        """Preset skill overrides must resolve __SPECKIT_COMMAND_*__ tokens (issue #2717).
+
+        ``_register_skills()`` previously ran only ``resolve_skill_placeholders()``,
+        so command cross-references leaked into SKILL.md as raw placeholders
+        instead of rendering as ``/speckit-<cmd>`` like the command layer.
+        """
+        self._write_init_options(project_dir, ai="claude")
+        skills_dir = project_dir / ".claude" / "skills"
+        self._create_skill(skills_dir, "speckit-specify")
+
+        preset_dir = self._create_command_preset(
+            temp_dir,
+            "cmdref-install",
+            "speckit.specify",
+            "Override specify",
+            "Run `__SPECKIT_COMMAND_SPECIFY__` then `__SPECKIT_COMMAND_PLAN__`.\n",
+        )
+
+        manager = PresetManager(project_dir)
+        manager.install_from_directory(preset_dir, "0.1.5")
+
+        content = (skills_dir / "speckit-specify" / "SKILL.md").read_text()
+        assert "__SPECKIT_COMMAND_" not in content, "raw command token leaked into SKILL.md"
+        # Claude's invoke_separator is "-", so tokens render as /speckit-<cmd>.
+        assert "/speckit-specify" in content
+        assert "/speckit-plan" in content
+
+    def test_restore_skill_resolves_command_refs(self, project_dir, temp_dir):
+        """Skill restore on preset removal must also resolve command tokens (issue #2717)."""
+        self._write_init_options(project_dir, ai="claude")
+        skills_dir = project_dir / ".claude" / "skills"
+        self._create_skill(skills_dir, "speckit-specify")
+
+        core_cmds = project_dir / ".specify" / "templates" / "commands"
+        core_cmds.mkdir(parents=True, exist_ok=True)
+        (core_cmds / "specify.md").write_text(
+            "---\ndescription: Core specify\n---\n\n"
+            "Then run `__SPECKIT_COMMAND_PLAN__`.\n"
+        )
+
+        preset_dir = self._create_command_preset(
+            temp_dir,
+            "cmdref-restore",
+            "speckit.specify",
+            "Override specify",
+            "Override body\n",
+        )
+        manager = PresetManager(project_dir)
+        manager.install_from_directory(preset_dir, "0.1.5")
+        manager.remove("cmdref-restore")
+
+        content = (skills_dir / "speckit-specify" / "SKILL.md").read_text()
+        assert "__SPECKIT_COMMAND_" not in content, "raw command token leaked on restore"
+        assert "/speckit-plan" in content
+
+    def test_reconcile_override_skill_resolves_command_refs(self, project_dir, temp_dir):
+        """Reconcile's project-override restore must resolve command tokens (issue #2717).
+
+        When a preset that overrode a command is removed and a project override
+        becomes the winning layer, ``_reconcile_skills`` rewrites the skill from
+        the override body — which must also render ``__SPECKIT_COMMAND_*__`` tokens.
+        """
+        self._write_init_options(project_dir, ai="claude")
+        skills_dir = project_dir / ".claude" / "skills"
+        self._create_skill(skills_dir, "speckit-specify")
+
+        # Project override wins once the preset is removed; its body carries a
+        # command cross-reference token. No core template exists for "specify",
+        # so the skill is restored exclusively via the reconcile override branch.
+        overrides_dir = project_dir / ".specify" / "templates" / "overrides"
+        overrides_dir.mkdir(parents=True, exist_ok=True)
+        (overrides_dir / "speckit.specify.md").write_text(
+            "---\ndescription: Override specify\n---\n\n"
+            "Then run `__SPECKIT_COMMAND_PLAN__`.\n"
+        )
+
+        preset_dir = self._create_command_preset(
+            temp_dir,
+            "cmdref-reconcile",
+            "speckit.specify",
+            "Preset specify",
+            "Preset body\n",
+        )
+        manager = PresetManager(project_dir)
+        manager.install_from_directory(preset_dir, "0.1.5")
+        manager.remove("cmdref-reconcile")
+
+        content = (skills_dir / "speckit-specify" / "SKILL.md").read_text()
+        assert "override:speckit.specify" in content, "skill should be restored from the project override"
+        assert "__SPECKIT_COMMAND_" not in content, "raw command token leaked on reconcile"
+        assert "/speckit-plan" in content
+
+    def test_extension_restore_resolves_command_refs(self, project_dir, temp_dir):
+        """Extension-backed skill restore must resolve command tokens (issue #2717).
+
+        When a preset override is removed and the skill is restored from an
+        extension command body, ``__SPECKIT_COMMAND_*__`` tokens in that body
+        must render as slash-command invocations like the core-template path.
+        """
+        self._write_init_options(project_dir, ai="claude")
+        skills_dir = project_dir / ".claude" / "skills"
+        self._create_skill(skills_dir, "speckit-fakeext-cmd", body="original extension skill")
+
+        extension_dir = project_dir / ".specify" / "extensions" / "fakeext"
+        (extension_dir / "commands").mkdir(parents=True, exist_ok=True)
+        (extension_dir / "commands" / "cmd.md").write_text(
+            "---\ndescription: Extension fakeext cmd\n---\n\n"
+            "Then run `__SPECKIT_COMMAND_PLAN__`.\n"
+        )
+        extension_manifest = {
+            "schema_version": "1.0",
+            "extension": {
+                "id": "fakeext",
+                "name": "Fake Extension",
+                "version": "1.0.0",
+                "description": "Test",
+            },
+            "requires": {"speckit_version": ">=0.1.0"},
+            "provides": {
+                "commands": [
+                    {
+                        "name": "speckit.fakeext.cmd",
+                        "file": "commands/cmd.md",
+                        "description": "Fake extension command",
+                    }
+                ]
+            },
+        }
+        with open(extension_dir / "extension.yml", "w") as f:
+            yaml.dump(extension_manifest, f)
+
+        preset_dir = self._create_command_preset(
+            temp_dir,
+            "cmdref-ext-restore",
+            "speckit.fakeext.cmd",
+            "Override fakeext cmd",
+            "Override body\n",
+        )
+        manager = PresetManager(project_dir)
+        manager.install_from_directory(preset_dir, "0.1.5")
+        manager.remove("cmdref-ext-restore")
+
+        content = (skills_dir / "speckit-fakeext-cmd" / "SKILL.md").read_text()
+        assert "source: extension:fakeext" in content, "skill should be restored from the extension"
+        assert "__SPECKIT_COMMAND_" not in content, "raw command token leaked on extension restore"
+        assert "/speckit-plan" in content
+
     def test_core_command_override_skill_uses_preset_command_description(self, project_dir, temp_dir):
         """Preset skill overrides for core commands should keep preset frontmatter descriptions."""
         self._write_init_options(project_dir, ai="claude")

tests/test_ps1_encoding.py

@@ -0,0 +1,54 @@
+"""Regression tests for PowerShell 5.1 compatibility (GitHub issue #2680).
+
+PowerShell 5.1 (built-in on Windows) defaults to the system's legacy encoding
+when reading .ps1 files.  Non-ASCII characters in UTF-8-encoded scripts cause
+parse errors because multi-byte sequences are misinterpreted as individual bytes.
+
+These tests ensure that all shipped .ps1 files remain ASCII-only so they work
+on both PowerShell 5.1 and 7+.
+"""
+
+from pathlib import Path
+
+import pytest
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+
+# All directories that contain shipped PowerShell scripts.
+_PS1_DIRS = [
+    REPO_ROOT / "scripts" / "powershell",
+    REPO_ROOT / "extensions" / "git" / "scripts" / "powershell",
+]
+
+
+def _collect_ps1_files():
+    """Yield all .ps1 files under the known script directories."""
+    for d in _PS1_DIRS:
+        if d.is_dir():
+            yield from sorted(d.rglob("*.ps1"))
+
+
+_PS1_FILES = list(_collect_ps1_files())
+
+
+@pytest.mark.parametrize("ps1_file", _PS1_FILES, ids=lambda p: str(p.relative_to(REPO_ROOT)))
+def test_ps1_file_is_ascii_only(ps1_file: Path):
+    """Every .ps1 file must contain only ASCII characters (PS 5.1 compat)."""
+    content = ps1_file.read_bytes()
+    non_ascii = [
+        (i + 1, byte)
+        for i, byte in enumerate(content)
+        if byte > 127
+    ]
+    assert not non_ascii, (
+        f"{ps1_file.relative_to(REPO_ROOT)} contains non-ASCII bytes "
+        f"(PowerShell 5.1 incompatible): "
+        f"first at byte offset {non_ascii[0][0]} (0x{non_ascii[0][1]:02x})"
+    )
+
+
+def test_ps1_files_discovered():
+    """Sanity check: at least the known script files are found."""
+    names = {p.name for p in _PS1_FILES}
+    assert "common.ps1" in names
+    assert "initialize-repo.ps1" in names

tests/test_setup_plan_no_overwrite.py

@@ -0,0 +1,216 @@
+"""Tests for setup-plan preserving existing plan.md (#2653)."""
+
+import json
+import os
+import shutil
+import subprocess
+from pathlib import Path
+
+import pytest
+
+from tests.conftest import requires_bash
+
+PROJECT_ROOT = Path(__file__).resolve().parent.parent
+COMMON_SH = PROJECT_ROOT / "scripts" / "bash" / "common.sh"
+SETUP_PLAN_SH = PROJECT_ROOT / "scripts" / "bash" / "setup-plan.sh"
+COMMON_PS = PROJECT_ROOT / "scripts" / "powershell" / "common.ps1"
+SETUP_PLAN_PS = PROJECT_ROOT / "scripts" / "powershell" / "setup-plan.ps1"
+PLAN_TEMPLATE = PROJECT_ROOT / "templates" / "plan-template.md"
+
+HAS_PWSH = shutil.which("pwsh") is not None
+_POWERSHELL = shutil.which("powershell.exe") or shutil.which("powershell")
+
+
+def _install_bash_scripts(repo: Path) -> None:
+    d = repo / ".specify" / "scripts" / "bash"
+    d.mkdir(parents=True, exist_ok=True)
+    shutil.copy(COMMON_SH, d / "common.sh")
+    shutil.copy(SETUP_PLAN_SH, d / "setup-plan.sh")
+
+
+def _install_ps_scripts(repo: Path) -> None:
+    d = repo / ".specify" / "scripts" / "powershell"
+    d.mkdir(parents=True, exist_ok=True)
+    shutil.copy(COMMON_PS, d / "common.ps1")
+    shutil.copy(SETUP_PLAN_PS, d / "setup-plan.ps1")
+
+
+def _minimal_templates(repo: Path) -> None:
+    tdir = repo / ".specify" / "templates"
+    tdir.mkdir(parents=True, exist_ok=True)
+    shutil.copy(PLAN_TEMPLATE, tdir / "plan-template.md")
+
+
+def _clean_env() -> dict[str, str]:
+    env = os.environ.copy()
+    for key in list(env):
+        if key.startswith("SPECIFY_"):
+            env.pop(key)
+    return env
+
+
+def _git_init(repo: Path) -> None:
+    subprocess.run(["git", "init", "-q"], cwd=repo, check=True)
+    subprocess.run(
+        ["git", "config", "user.email", "test@example.com"], cwd=repo, check=True
+    )
+    subprocess.run(["git", "config", "user.name", "Test User"], cwd=repo, check=True)
+    subprocess.run(
+        ["git", "commit", "--allow-empty", "-m", "init", "-q"], cwd=repo, check=True
+    )
+
+
+@pytest.fixture
+def plan_repo(tmp_path: Path) -> Path:
+    repo = tmp_path / "proj"
+    repo.mkdir()
+    _git_init(repo)
+    subprocess.run(
+        ["git", "checkout", "-q", "-b", "001-my-feature"],
+        cwd=repo,
+        check=True,
+    )
+    (repo / ".specify").mkdir()
+    _minimal_templates(repo)
+    _install_bash_scripts(repo)
+    _install_ps_scripts(repo)
+    return repo
+
+
+# ── Bash tests ────────────────────────────────────────────────────────────
+
+
+@requires_bash
+def test_setup_plan_creates_plan_when_missing(plan_repo: Path) -> None:
+    """First run must create plan.md from the template."""
+    script = plan_repo / ".specify" / "scripts" / "bash" / "setup-plan.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    plan_path = Path(data["IMPL_PLAN"])
+    assert plan_path.is_file()
+    # Template content should be present
+    content = plan_path.read_text(encoding="utf-8")
+    assert len(content) > 0
+
+
+@requires_bash
+def test_setup_plan_preserves_existing_plan(plan_repo: Path) -> None:
+    """Rerun must not overwrite an existing plan.md."""
+    feat = plan_repo / "specs" / "001-my-feature"
+    feat.mkdir(parents=True)
+    existing_content = "# My carefully authored plan\n\nDo not overwrite me.\n"
+    (feat / "plan.md").write_text(existing_content, encoding="utf-8")
+
+    script = plan_repo / ".specify" / "scripts" / "bash" / "setup-plan.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    # Plan must be unchanged
+    assert (feat / "plan.md").read_text(encoding="utf-8") == existing_content
+
+
+@requires_bash
+def test_setup_plan_skip_message_on_stderr_in_json_mode(plan_repo: Path) -> None:
+    """In --json mode, status messages must go to stderr, not stdout."""
+    feat = plan_repo / "specs" / "001-my-feature"
+    feat.mkdir(parents=True)
+    (feat / "plan.md").write_text("# existing\n", encoding="utf-8")
+
+    script = plan_repo / ".specify" / "scripts" / "bash" / "setup-plan.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    # stdout must be valid JSON (no status messages mixed in)
+    data = json.loads(result.stdout)
+    assert "IMPL_PLAN" in data
+    # The skip message should be on stderr
+    assert "already exists" in result.stderr
+
+
+@requires_bash
+def test_setup_plan_json_parseable_on_first_run(plan_repo: Path) -> None:
+    """In --json mode, first-run stdout must be parseable JSON (no status on stdout)."""
+    script = plan_repo / ".specify" / "scripts" / "bash" / "setup-plan.sh"
+    result = subprocess.run(
+        ["bash", str(script), "--json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    assert "IMPL_PLAN" in data
+    assert "Copied plan template" in result.stderr
+
+
+# ── PowerShell tests ──────────────────────────────────────────────────────
+
+
+@pytest.mark.skipif(not (HAS_PWSH or _POWERSHELL), reason="no PowerShell available")
+def test_ps_setup_plan_creates_plan_when_missing(plan_repo: Path) -> None:
+    """First run must create plan.md from the template."""
+    script = plan_repo / ".specify" / "scripts" / "powershell" / "setup-plan.ps1"
+    exe = "pwsh" if HAS_PWSH else _POWERSHELL
+    result = subprocess.run(
+        [exe, "-NoProfile", "-File", str(script), "-Json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    data = json.loads(result.stdout)
+    plan_path = Path(data["IMPL_PLAN"])
+    assert plan_path.is_file()
+    content = plan_path.read_text(encoding="utf-8")
+    assert len(content) > 0
+
+
+@pytest.mark.skipif(not (HAS_PWSH or _POWERSHELL), reason="no PowerShell available")
+def test_ps_setup_plan_preserves_existing_plan(plan_repo: Path) -> None:
+    """Rerun must not overwrite an existing plan.md."""
+    feat = plan_repo / "specs" / "001-my-feature"
+    feat.mkdir(parents=True)
+    existing_content = "# My carefully authored plan\n\nDo not overwrite me.\n"
+    (feat / "plan.md").write_text(existing_content, encoding="utf-8")
+
+    script = plan_repo / ".specify" / "scripts" / "powershell" / "setup-plan.ps1"
+    exe = "pwsh" if HAS_PWSH else _POWERSHELL
+    result = subprocess.run(
+        [exe, "-NoProfile", "-File", str(script), "-Json"],
+        cwd=plan_repo,
+        capture_output=True,
+        text=True,
+        check=False,
+        env=_clean_env(),
+    )
+    assert result.returncode == 0, result.stderr
+    assert (feat / "plan.md").read_text(encoding="utf-8") == existing_content
+    # stdout must be valid JSON (no status messages mixed in)
+    data = json.loads(result.stdout)
+    assert "IMPL_PLAN" in data
+    # The skip message should be on stderr
+    assert "already exists" in result.stderr

tests/test_workflows.py

@@ -13,6 +13,7 @@ Covers:
 from __future__ import annotations
 
 import json
+import os
 import shutil
 import tempfile
 from pathlib import Path
@@ -332,6 +333,44 @@ class TestExpressions:
         result = evaluate_expression("{{ steps.tasks.output.task_list[0].file }}", ctx)
         assert result == "a.md"
 
+    def test_context_run_id_resolves(self):
+        """``{{ context.run_id }}`` resolves to ``StepContext.run_id``.
+
+        Locks the contract from issue #2590: workflow templates can
+        reference the engine-assigned run id for telemetry, artifact
+        metadata, or per-run scratch isolation.
+        """
+        from specify_cli.workflows.expressions import evaluate_expression
+        from specify_cli.workflows.base import StepContext
+
+        ctx = StepContext(run_id="a1b2c3d4")
+        assert evaluate_expression("{{ context.run_id }}", ctx) == "a1b2c3d4"
+
+    def test_context_run_id_defaults_to_empty_when_unset(self):
+        """``{{ context.run_id }}`` resolves to ``""`` when no run is
+        active (dry-run, validation, ad-hoc evaluator usage) rather
+        than raising — workflows referencing the variable never error
+        outside a run context.
+        """
+        from specify_cli.workflows.expressions import evaluate_expression
+        from specify_cli.workflows.base import StepContext
+
+        # No run_id set on the context.
+        ctx = StepContext()
+        assert evaluate_expression("{{ context.run_id }}", ctx) == ""
+
+    def test_context_run_id_string_interpolation(self):
+        """Run id interpolates inside a larger template string — the
+        common pattern for stamping shell commands and artifact paths
+        with the run id.
+        """
+        from specify_cli.workflows.expressions import evaluate_expression
+        from specify_cli.workflows.base import StepContext
+
+        ctx = StepContext(run_id="deadbeef")
+        result = evaluate_expression("RUN_ID={{ context.run_id }}", ctx)
+        assert result == "RUN_ID=deadbeef"
+
 
 # ===== Integration Dispatch Tests =====
 
@@ -373,7 +412,8 @@ class TestBuildExecArgs:
         from specify_cli.integrations.copilot import CopilotIntegration
         impl = CopilotIntegration()
         args = impl.build_exec_args("do stuff", model="claude-sonnet-4-20250514")
-        assert args[0] == "copilot"
+        expected_exec = "copilot.cmd" if os.name == "nt" else "copilot"
+        assert args[0] == expected_exec
         assert "-p" in args
         assert "--yolo" in args
         assert "--model" in args
@@ -463,6 +503,7 @@ class TestCommandStep:
         assert any("missing 'command'" in e for e in errors)
 
     def test_step_override_integration(self):
+        from unittest.mock import patch
         from specify_cli.workflows.steps.command import CommandStep
         from specify_cli.workflows.base import StepContext
 
@@ -474,7 +515,8 @@ class TestCommandStep:
             "integration": "gemini",
             "input": {},
         }
-        result = step.execute(config, ctx)
+        with patch("specify_cli.workflows.steps.command.shutil.which", return_value=None):
+            result = step.execute(config, ctx)
         assert result.output["integration"] == "gemini"
 
     def test_step_override_model(self):
@@ -626,6 +668,7 @@ class TestPromptStep:
         assert result.output["dispatched"] is False
 
     def test_execute_with_step_integration(self):
+        from unittest.mock import patch
         from specify_cli.workflows.steps.prompt import PromptStep
         from specify_cli.workflows.base import StepContext
 
@@ -637,10 +680,12 @@ class TestPromptStep:
             "prompt": "Summarize the codebase",
             "integration": "gemini",
         }
-        result = step.execute(config, ctx)
+        with patch("specify_cli.workflows.steps.prompt.shutil.which", return_value=None):
+            result = step.execute(config, ctx)
         assert result.output["integration"] == "gemini"
 
     def test_execute_with_model(self):
+        from unittest.mock import patch
         from specify_cli.workflows.steps.prompt import PromptStep
         from specify_cli.workflows.base import StepContext
 
@@ -652,7 +697,8 @@ class TestPromptStep:
             "prompt": "hello",
             "model": "opus-4",
         }
-        result = step.execute(config, ctx)
+        with patch("specify_cli.workflows.steps.prompt.shutil.which", return_value=None):
+            result = step.execute(config, ctx)
         assert result.output["model"] == "opus-4"
 
     def test_dispatch_with_mock_cli(self, tmp_path):
@@ -1883,6 +1929,451 @@ steps:
         errors = validate_workflow(definition)
         assert any("invalid default" in e for e in errors), errors
 
+    def test_while_loop_condition_reads_latest_iteration(self, project_dir):
+        """Regression: while-loop condition must see updated step output
+        from the most recent iteration, not stale iteration-0 data.
+
+        See https://github.com/github/spec-kit/issues/2592
+        """
+        from specify_cli.workflows.engine import WorkflowEngine, WorkflowDefinition
+        from specify_cli.workflows.base import RunStatus
+
+        # Shell step echoes a counter via a file.
+        # Condition: exit_code != 0 means "keep looping" — but a non-zero
+        # exit code would mark the step FAILED and abort the run, so we
+        # use stdout-based comparison instead.
+        #
+        # Iteration 0: counter=1, echoes "1" → not "done" → loop continues
+        # Iteration 1: counter=2, echoes "done" → condition false → stop
+        # Without the fix, condition always reads iteration-0 stdout,
+        # so the loop runs all max_iterations.
+        import sys
+
+        counter_file = project_dir / ".counter"
+        counter_file.write_text("0", encoding="utf-8")
+        py = sys.executable
+        script_file = project_dir / "_tick.py"
+        script_file.write_text(
+            f"import pathlib; p = pathlib.Path(r'{counter_file}')\n"
+            "n = int(p.read_text()) + 1; p.write_text(str(n))\n"
+            "print('done' if n >= 2 else str(n), end='')\n",
+            encoding="utf-8",
+        )
+
+        yaml_str = f"""
+schema_version: "1.0"
+workflow:
+  id: "while-condition-update"
+  name: "While Condition Update"
+  version: "1.0.0"
+steps:
+  - id: retry-loop
+    type: while
+    condition: "{{{{ 'done' not in steps.attempt.output.stdout }}}}"
+    max_iterations: 5
+    steps:
+      - id: attempt
+        type: shell
+        run: '"{py}" "{script_file}"'
+"""
+        definition = WorkflowDefinition.from_string(yaml_str)
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        # The unprefixed key should reflect the latest iteration's result.
+        assert state.step_results["attempt"]["output"]["stdout"] == "done"
+        # Namespaced iteration-1 result should also exist.
+        assert "retry-loop:attempt:1" in state.step_results
+        # Counter should be 2 (iteration 0 + iteration 1), not 5.
+        assert counter_file.read_text(encoding="utf-8").strip() == "2"
+
+    def test_do_while_loop_condition_reads_latest_iteration(self, project_dir):
+        """Regression: do-while loop condition must also see updated output.
+
+        See https://github.com/github/spec-kit/issues/2592
+        """
+        from specify_cli.workflows.engine import WorkflowEngine, WorkflowDefinition
+        from specify_cli.workflows.base import RunStatus
+
+        import sys
+
+        counter_file = project_dir / ".counter"
+        counter_file.write_text("0", encoding="utf-8")
+        py = sys.executable
+        script_file = project_dir / "_tick.py"
+        script_file.write_text(
+            f"import pathlib; p = pathlib.Path(r'{counter_file}')\n"
+            "n = int(p.read_text()) + 1; p.write_text(str(n))\n"
+            "print('done' if n >= 2 else str(n), end='')\n",
+            encoding="utf-8",
+        )
+
+        yaml_str = f"""
+schema_version: "1.0"
+workflow:
+  id: "do-while-condition-update"
+  name: "Do While Condition Update"
+  version: "1.0.0"
+steps:
+  - id: retry-loop
+    type: do-while
+    condition: "{{{{ 'done' not in steps.attempt.output.stdout }}}}"
+    max_iterations: 5
+    steps:
+      - id: attempt
+        type: shell
+        run: '"{py}" "{script_file}"'
+"""
+        definition = WorkflowDefinition.from_string(yaml_str)
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        assert state.step_results["attempt"]["output"]["stdout"] == "done"
+        assert counter_file.read_text(encoding="utf-8").strip() == "2"
+
+    def test_while_loop_runs_to_max_when_condition_stays_true(self, project_dir):
+        """While loop must still run to max_iterations when the condition
+        never becomes false — copy-back must not break this path.
+
+        See https://github.com/github/spec-kit/issues/2592
+        """
+        from specify_cli.workflows.engine import WorkflowEngine, WorkflowDefinition
+        from specify_cli.workflows.base import RunStatus
+
+        import sys
+
+        counter_file = project_dir / ".counter"
+        counter_file.write_text("0", encoding="utf-8")
+        py = sys.executable
+        script_file = project_dir / "_tick.py"
+        script_file.write_text(
+            f"import pathlib; p = pathlib.Path(r'{counter_file}')\n"
+            "n = int(p.read_text()) + 1; p.write_text(str(n))\n"
+            "print('pending', end='')\n",
+            encoding="utf-8",
+        )
+
+        yaml_str = f"""
+schema_version: "1.0"
+workflow:
+  id: "while-max-iterations"
+  name: "While Max Iterations"
+  version: "1.0.0"
+steps:
+  - id: retry-loop
+    type: while
+    condition: "{{{{ 'done' not in steps.tick.output.stdout }}}}"
+    max_iterations: 3
+    steps:
+      - id: tick
+        type: shell
+        run: '"{py}" "{script_file}"'
+"""
+        definition = WorkflowDefinition.from_string(yaml_str)
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        # All 3 iterations ran (iteration 0 + 2 loop iterations).
+        assert counter_file.read_text(encoding="utf-8").strip() == "3"
+        # Unprefixed key holds the last iteration's result.
+        assert state.step_results["tick"]["output"]["stdout"] == "pending"
+        # Namespaced keys for loop iterations exist.
+        assert "retry-loop:tick:1" in state.step_results
+        assert "retry-loop:tick:2" in state.step_results
+
+    def test_do_while_loop_runs_to_max_when_condition_stays_true(self, project_dir):
+        """Do-while loop must still run to max_iterations when the condition
+        never becomes false.
+
+        See https://github.com/github/spec-kit/issues/2592
+        """
+        from specify_cli.workflows.engine import WorkflowEngine, WorkflowDefinition
+        from specify_cli.workflows.base import RunStatus
+
+        import sys
+
+        counter_file = project_dir / ".counter"
+        counter_file.write_text("0", encoding="utf-8")
+        py = sys.executable
+        script_file = project_dir / "_tick.py"
+        script_file.write_text(
+            f"import pathlib; p = pathlib.Path(r'{counter_file}')\n"
+            "n = int(p.read_text()) + 1; p.write_text(str(n))\n"
+            "print('pending', end='')\n",
+            encoding="utf-8",
+        )
+
+        yaml_str = f"""
+schema_version: "1.0"
+workflow:
+  id: "do-while-max-iterations"
+  name: "Do While Max Iterations"
+  version: "1.0.0"
+steps:
+  - id: retry-loop
+    type: do-while
+    condition: "{{{{ 'done' not in steps.tick.output.stdout }}}}"
+    max_iterations: 3
+    steps:
+      - id: tick
+        type: shell
+        run: '"{py}" "{script_file}"'
+"""
+        definition = WorkflowDefinition.from_string(yaml_str)
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        assert counter_file.read_text(encoding="utf-8").strip() == "3"
+        assert state.step_results["tick"]["output"]["stdout"] == "pending"
+
+    def test_while_loop_multi_step_body_inter_step_refs(self, project_dir):
+        """Multi-step loop body: step B must see step A's output from the
+        current iteration, not a stale previous one.
+
+        See https://github.com/github/spec-kit/issues/2592
+        """
+        from specify_cli.workflows.engine import WorkflowEngine, WorkflowDefinition
+        from specify_cli.workflows.base import RunStatus
+
+        import sys
+
+        counter_file = project_dir / ".counter"
+        counter_file.write_text("0", encoding="utf-8")
+        py = sys.executable
+
+        # Step A: increments counter file, echoes the value.
+        step_a_file = project_dir / "_step_a.py"
+        step_a_file.write_text(
+            f"import pathlib; p = pathlib.Path(r'{counter_file}')\n"
+            "n = int(p.read_text()) + 1; p.write_text(str(n))\n"
+            "print(str(n), end='')\n",
+            encoding="utf-8",
+        )
+
+        # Step B uses {{ steps.step-a.output.stdout }} expression
+        # substitution in its run command so the engine resolves the
+        # aliased unprefixed key — this is the real inter-step test.
+        yaml_str = f"""
+schema_version: "1.0"
+workflow:
+  id: "while-multi-step"
+  name: "While Multi Step"
+  version: "1.0.0"
+steps:
+  - id: retry-loop
+    type: while
+    condition: "{{{{ 'done' not in steps.step-a.output.stdout }}}}"
+    max_iterations: 3
+    steps:
+      - id: step-a
+        type: shell
+        run: '"{py}" "{step_a_file}"'
+      - id: step-b
+        type: shell
+        run: "echo b-saw-{{{{ steps.step-a.output.stdout }}}}"
+"""
+        definition = WorkflowDefinition.from_string(yaml_str)
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        # Both unprefixed keys reflect the latest iteration's results.
+        assert state.step_results["step-a"]["output"]["stdout"] == "3"
+        # Step B saw step A's output via expression substitution.
+        assert "b-saw-3" in state.step_results["step-b"]["output"]["stdout"]
+        # Namespaced keys exist for loop iterations.
+        assert "retry-loop:step-a:1" in state.step_results
+        assert "retry-loop:step-b:1" in state.step_results
+        assert "retry-loop:step-a:2" in state.step_results
+        assert "retry-loop:step-b:2" in state.step_results
+
+
+# ===== context.run_id Tests =====
+#
+# End-to-end coverage for the `{{ context.run_id }}` template
+# variable introduced in issue #2590. Locks resolution inside the
+# three step types the acceptance criteria called out — shell `run:`,
+# command `input.args:`, and switch `expression:` — plus the
+# "workflow doesn't reference it" backward-compat path.
+
+
+class TestContextRunId:
+    """End-to-end tests for `{{ context.run_id }}` in workflow YAML."""
+
+    def test_shell_run_resolves_run_id(self, project_dir):
+        """`run: "echo {{ context.run_id }}"` substitutes the
+        engine-assigned run id into the spawned shell, and the
+        same value appears on `state.run_id`.
+        """
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "stamp-run-id"
+  name: "Stamp Run Id"
+  version: "1.0.0"
+steps:
+  - id: stamp
+    type: shell
+    run: "echo RUN_ID={{ context.run_id }}"
+""")
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition, run_id="abc12345")
+
+        assert state.run_id == "abc12345"
+        stdout = state.step_results["stamp"]["output"]["stdout"]
+        assert stdout.strip() == "RUN_ID=abc12345"
+
+    def test_command_input_args_resolves_run_id(self, project_dir):
+        """`input.args: "{{ context.run_id }}"` is resolved by
+        `CommandStep` and recorded in step output, even when CLI
+        dispatch is unavailable (no integration installed). Covers
+        the artifact-metadata use case from the issue.
+        """
+        from unittest.mock import patch
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "command-stamp"
+  name: "Command Stamp"
+  version: "1.0.0"
+  integration: claude
+steps:
+  - id: tag-artifact
+    command: speckit.specify
+    input:
+      args: "{{ context.run_id }}"
+""")
+        engine = WorkflowEngine(project_dir)
+        with patch(
+            "specify_cli.workflows.steps.command.shutil.which",
+            return_value=None,
+        ):
+            state = engine.execute(definition, run_id="cafef00d")
+
+        # Even when dispatch fails (no CLI), the resolved input is
+        # recorded so downstream observers see the run id in artifact
+        # metadata.
+        assert state.step_results["tag-artifact"]["output"]["input"]["args"] == "cafef00d"
+
+    def test_switch_expression_matches_on_run_id(self, project_dir):
+        """`switch` over `{{ context.run_id }}` matches against case
+        keys, and the nested branch can ALSO reference
+        `{{ context.run_id }}`. Demonstrates the run id is a
+        first-class value in the expression engine (not just a
+        string-interpolation token) AND that it propagates into
+        nested step execution via the recursive `_execute_steps`
+        traversal.
+        """
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+        from specify_cli.workflows.base import RunStatus
+
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "switch-on-run-id"
+  name: "Switch On Run Id"
+  version: "1.0.0"
+steps:
+  - id: route
+    type: switch
+    expression: "{{ context.run_id }}"
+    cases:
+      target-run:
+        - id: matched-branch
+          type: shell
+          run: "echo nested-run-id={{ context.run_id }}"
+    default:
+      - id: default-branch
+        type: shell
+        run: "echo defaulted"
+""")
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition, run_id="target-run")
+
+        assert state.status == RunStatus.COMPLETED
+        assert state.step_results["route"]["output"]["matched_case"] == "target-run"
+        assert "matched-branch" in state.step_results
+        assert "default-branch" not in state.step_results
+        # The nested branch sees the same run id — propagation through
+        # recursive `_execute_steps` is intact.
+        nested_stdout = state.step_results["matched-branch"]["output"]["stdout"]
+        assert nested_stdout.strip() == "nested-run-id=target-run"
+
+    def test_workflow_without_context_reference_unchanged(self, project_dir):
+        """Workflows that do not reference `{{ context.run_id }}`
+        continue to run exactly as before. Locks the byte-equivalent
+        default required by the issue's acceptance criteria.
+        """
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+        from specify_cli.workflows.base import RunStatus
+
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "no-context-ref"
+  name: "No Context Ref"
+  version: "1.0.0"
+steps:
+  - id: only-step
+    type: shell
+    run: "echo hello"
+""")
+        engine = WorkflowEngine(project_dir)
+        state = engine.execute(definition)
+
+        assert state.status == RunStatus.COMPLETED
+        assert state.step_results["only-step"]["output"]["stdout"].strip() == "hello"
+
+    def test_run_id_uses_speckit_workflow_run_id_env_override(self, project_dir, monkeypatch):
+        """When no run_id argument is provided, SPECKIT_WORKFLOW_RUN_ID overrides the auto-generated run ID."""
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+
+        monkeypatch.setenv("SPECKIT_WORKFLOW_RUN_ID", "env-run-123")
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "env-run-id"
+  name: "Env Run Id"
+  version: "1.0.0"
+steps:
+  - id: stamp
+    type: shell
+    run: "echo {{ context.run_id }}"
+""")
+        state = WorkflowEngine(project_dir).execute(definition)
+
+        assert state.run_id == "env-run-123"
+        assert state.step_results["stamp"]["output"]["stdout"].strip() == "env-run-123"
+
+    def test_run_id_arg_takes_precedence_over_env_override(self, project_dir, monkeypatch):
+        """Explicit run_id keeps existing precedence over SPECKIT_WORKFLOW_RUN_ID."""
+        from specify_cli.workflows.engine import WorkflowDefinition, WorkflowEngine
+
+        monkeypatch.setenv("SPECKIT_WORKFLOW_RUN_ID", "env-run-123")
+        definition = WorkflowDefinition.from_string("""
+schema_version: "1.0"
+workflow:
+  id: "explicit-run-id"
+  name: "Explicit Run Id"
+  version: "1.0.0"
+steps:
+  - id: stamp
+    type: shell
+    run: "echo {{ context.run_id }}"
+""")
+        state = WorkflowEngine(project_dir).execute(definition, run_id="explicit-456")
+
+        assert state.run_id == "explicit-456"
+        assert state.step_results["stamp"]["output"]["stdout"].strip() == "explicit-456"
+
 
 # ===== State Persistence Tests =====
 

workflows/README.md

@@ -239,6 +239,33 @@ message: "{{ status | default('pending') }}"
 
 Supported filters: `default`, `join`, `contains`, `map`.
 
+### Runtime Context
+
+`{{ context.* }}` exposes engine-managed runtime metadata for the
+current run:
+
+| Variable | Description |
+|----------|-------------|
+| `context.run_id` | The current workflow run id (the same value Spec Kit prints as `Run ID:` at the end of `workflow run`). Auto-generated runs are 8-character hex from `uuid4`; operator-supplied ids may be any alphanumeric string with hyphens or underscores. Empty string outside a run context. |
+
+```yaml
+# Stamp telemetry events with the run id for cross-system join.
+- id: emit-event
+  type: shell
+  run: 'echo "{\"run_id\":\"{{ context.run_id }}\",\"event\":\"started\"}" >> events.jsonl'
+
+# Per-run scratch directory.
+- id: prep-scratch
+  type: shell
+  run: 'mkdir -p /tmp/run-{{ context.run_id }}'
+
+# Pass run id into a command for artifact metadata.
+- id: tag-artifact
+  command: speckit.specify
+  input:
+    args: "{{ context.run_id }}"
+```
+
 ## Input Types
 
 Workflow inputs are type-checked and coerced from CLI string values: