tweak

Co-authored-by: Copilot <copilot@github.com>

.changeset/config.json

@@ -1,11 +1,11 @@
 {
 	"$schema": "https://unpkg.com/@changesets/config@3.1.1/schema.json",
-	"changelog": ["@svitejs/changesets-changelog-github-compact", { "repo": "sveltejs/mcp" }],
+	"changelog": ["@svitejs/changesets-changelog-github-compact", { "repo": "sveltejs/ai-tools" }],
 	"commit": false,
 	"fixed": [],
 	"linked": [],
 	"access": "public",
 	"baseBranch": "main",
 	"updateInternalDependencies": "patch",
-	"ignore": ["!@sveltejs/mcp"]
+	"ignore": ["!@sveltejs/mcp", "!@sveltejs/opencode"]
 }

.changeset/salty-jokes-crash.md

@@ -0,0 +1,5 @@
+---
+'@sveltejs/mcp': patch
+---
+
+fix: handle non call expressions passed to `is_rune`

.changeset/small-actors-notice.md

@@ -0,0 +1,5 @@
+---
+'@sveltejs/mcp': patch
+---
+
+chore: remove db requirement

.claude-plugin/marketplace.json

@@ -0,0 +1,22 @@
+{
+	"name": "svelte",
+	"owner": {
+		"name": "Svelte"
+	},
+	"plugins": [
+		{
+			"name": "svelte",
+			"source": "./plugins/claude/svelte",
+			"description": "A plugin for all things Svelte development, MCP, skills, and more.",
+			"lspServers": {
+				"svelte": {
+					"command": "svelteserver",
+					"args": ["--stdio"],
+					"extensionToLanguage": {
+						".svelte": "svelte"
+					}
+				}
+			}
+		}
+	]
+}

.cursor-plugin/marketplace.json

@@ -0,0 +1,13 @@
+{
+	"name": "svelte",
+	"owner": {
+		"name": "Svelte"
+	},
+	"plugins": [
+		{
+			"name": "svelte",
+			"source": "./plugins/cursor/svelte",
+			"description": "A plugin for all things Svelte development, MCP, skills, and more."
+		}
+	]
+}

.github/workflows/check.yml

@@ -13,17 +13,17 @@ jobs:
 
     steps:
       - name: Checkout Repository
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
 
       - name: Setup pnpm
-        uses: pnpm/action-setup@v4
+        uses: pnpm/action-setup@v5
         with:
-          version: 10.19.0
+          version: 10.28.2
 
       - name: Setup Node.js
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
-          node-version: '22'
+          node-version: '24'
           cache: 'pnpm'
 
       - name: Install dependencies
@@ -32,6 +32,4 @@ jobs:
       - name: Run type check
         run: pnpm run check
         env:
-          DATABASE_URL: file:test.db
-          DATABASE_TOKEN: dummy-key
           VOYAGE_API_KEY: dummy-key

.github/workflows/lint.yml

@@ -13,17 +13,17 @@ jobs:
 
     steps:
       - name: Checkout Repository
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
 
       - name: Setup pnpm
-        uses: pnpm/action-setup@v4
+        uses: pnpm/action-setup@v5
         with:
-          version: 10.19.0
+          version: 10.28.2
 
       - name: Setup Node.js
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
-          node-version: '22'
+          node-version: '24'
           cache: 'pnpm'
 
       - name: Install dependencies
@@ -32,6 +32,4 @@ jobs:
       - name: Run linting
         run: pnpm run lint
         env:
-          DATABASE_URL: file:test.db
           VOYAGE_API_KEY: dummy-key
-          DATABASE_TOKEN: dummy-key

.github/workflows/publish-any-commit.yml

@@ -0,0 +1,23 @@
+name: Publish Any Commit
+on: [push, pull_request]
+
+permissions: {}
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v6
+
+      - run: corepack enable
+      - uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          cache: 'pnpm'
+
+      - name: Install dependencies
+        run: pnpm install
+
+      - run: pnpm dlx pkg-pr-new publish --compact './packages/mcp-stdio' './packages/opencode' --pnpm

.github/workflows/publish-mcp.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: checkout
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
 
       - name: Publish to MCP Registry
         working-directory: packages/mcp-stdio
@@ -21,11 +21,11 @@ jobs:
           MCP_KEY: ${{ secrets.MCP_KEY }}
         run: |
           NAME=mcp-publisher_$(uname -s | tr '[:upper:]' '[:lower:]')_$(uname -m | sed 's/x86_64/amd64/;s/aarch64/arm64/').tar.gz
-          # Download MCP Publisher pinned to v1.3.3 using latest https for security and save it to a file named mcp-publisher.tar.gz
-          curl --proto '=https' --proto-redir '=https' --tlsv1.2 -fL "https://github.com/modelcontextprotocol/registry/releases/download/v1.3.3/$NAME" -O
+          # Download MCP Publisher pinned to v1.4.0 using latest https for security and save it to a file named mcp-publisher.tar.gz
+          curl --proto '=https' --proto-redir '=https' --tlsv1.2 -fL "https://github.com/modelcontextprotocol/registry/releases/download/v1.4.0/$NAME" -O
 
           # Verify the SHA256 checksum of the downloaded file
-          sha256sum --ignore-missing -c ./checksums/registry_1.3.3_checksums.txt
+          sha256sum --ignore-missing -c ./checksums/registry_1.4.0_checksums.txt
 
           # Extract the tarball
           mkdir tmp

.github/workflows/release-svelte-skill.yml

@@ -0,0 +1,66 @@
+name: Release Svelte Skills
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - 'tools/skills/**'
+  workflow_dispatch:
+
+permissions: {}
+
+jobs:
+  detect-skills:
+    permissions:
+      contents: read
+    if: github.repository == 'sveltejs/ai-tools'
+    name: Detect changed skills
+    runs-on: ubuntu-latest
+    outputs:
+      skills: ${{ steps.find-skills.outputs.skills }}
+    steps:
+      - name: checkout
+        uses: actions/checkout@v6
+
+      - name: Find all skills
+        id: find-skills
+        run: |
+          skills=$(ls -d tools/skills/*/ | xargs -I {} basename {} | jq -R -s -c 'split("\n") | map(select(length > 0))')
+          echo "skills=$skills" >> $GITHUB_OUTPUT
+
+  release:
+    needs: detect-skills
+    if: needs.detect-skills.outputs.skills != '[]'
+    permissions:
+      contents: write
+    name: Release ${{ matrix.skill }}
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        skill: ${{ fromJson(needs.detect-skills.outputs.skills) }}
+    steps:
+      - name: checkout
+        uses: actions/checkout@v6
+
+      - name: Get version from date
+        id: version
+        run: echo "version=$(date +'%Y.%m.%d-%H%M%S')" >> $GITHUB_OUTPUT
+
+      - name: Create zip
+        run: |
+          cd tools/skills
+          zip -r ${{ matrix.skill }}.zip ${{ matrix.skill }}/
+
+      - name: Create Release
+        uses: softprops/action-gh-release@v2
+        with:
+          tag_name: ${{ matrix.skill }}-v${{ steps.version.outputs.version }}
+          name: ${{ matrix.skill }} v${{ steps.version.outputs.version }}
+          body: |
+            Automated release of the ${{ matrix.skill }} skill.
+
+            This release was triggered by changes to the `tools/skills/${{ matrix.skill }}/` directory.
+          files: tools/skills/${{ matrix.skill }}.zip
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/release.yml

@@ -13,11 +13,11 @@ jobs:
       id-token: write # OpenID Connect token needed for provenance
       pull-requests: write # to create pull request (changesets/action)
     # prevents this action from running on forks
-    if: github.repository == 'sveltejs/mcp'
+    if: github.repository == 'sveltejs/ai-tools'
     name: Release
     runs-on: ${{ matrix.os }}
     outputs:
-      published: ${{ steps.changesets.outputs.published }}
+      publishedPackages: ${{ steps.changesets.outputs.publishedPackages }}
     strategy:
       matrix:
         # pseudo-matrix for convenience, NEVER use more than a single combination
@@ -25,11 +25,11 @@ jobs:
         os: [ubuntu-latest]
     steps:
       - name: checkout
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
         with:
           # This makes Actions fetch all Git history so that Changesets can generate changelogs with the correct commits
           fetch-depth: 0
-      - uses: actions/setup-node@v6
+      - uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
           node-version: ${{ matrix.node }}
           package-manager-cache: false # pnpm is not installed yet
@@ -39,10 +39,11 @@ jobs:
           PNPM_VER=$(jq -r '.packageManager | if .[0:5] == "pnpm@" then .[5:] else "packageManager in package.json does not start with pnpm@\n" | halt_error(1)  end' package.json)
           echo installing pnpm version $PNPM_VER
           npm i -g pnpm@$PNPM_VER
-      - uses: actions/setup-node@v6
+      - uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
           node-version: ${{ matrix.node }}
           package-manager-cache: true # caches pnpm via packageManager field in package.json
+          cache: 'pnpm'
       - name: install
         run: pnpm install --frozen-lockfile --prefer-offline --ignore-scripts
       - name: build
@@ -51,7 +52,7 @@ jobs:
       - name: Create Release Pull Request or Publish to npm
         id: changesets
         # pinned for security, always review third party action code before updating
-        uses: changesets/action@e0145edc7d9d8679003495b11f87bd8ef63c0cba # v1.5.3
+        uses: changesets/action@c48e67d110a68bc90ccf1098e9646092baacaa87 # v1.6.0
         with:
           # This expects you to have a script called changeset:version version that calls changeset version and updated what it needs to be updated
           version: pnpm changeset:version
@@ -63,7 +64,7 @@ jobs:
 
   publish-mcp:
     needs: release
-    if: needs.release.outputs.published == 'true'
+    if: contains(needs.release.outputs.publishedPackages, '"@sveltejs/mcp"')
     uses: ./.github/workflows/publish-mcp.yml
     secrets:
       MCP_KEY: ${{ secrets.MCP_KEY }}

.github/workflows/sync-docs-skills.yml

@@ -0,0 +1,116 @@
+name: Sync Skills
+
+on:
+  workflow_dispatch:
+
+permissions:
+  contents: write
+  pull-requests: write
+  actions: write
+
+jobs:
+  sync-skills:
+    if: github.repository == 'sveltejs/ai-tools'
+    name: Sync skills from svelte.dev
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Setup Node.js
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: false # pnpm is not installed yet
+
+      - name: Install pnpm
+        shell: bash
+        run: |
+          PNPM_VER=$(jq -r '.packageManager | if .[0:5] == "pnpm@" then .[5:] else "packageManager in package.json does not start with pnpm@\n" | halt_error(1) end' package.json)
+          echo installing pnpm version "$PNPM_VER"
+          npm i -g "pnpm@$PNPM_VER"
+
+      - name: Setup Node.js with pnpm cache
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: true # caches pnpm via packageManager field in package.json
+          cache: 'pnpm'
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile --prefer-offline --ignore-scripts
+
+      - name: Clone svelte.dev
+        run: git clone --depth 2 https://github.com/sveltejs/svelte.dev.git "${{ runner.temp }}/svelte.dev"
+
+      - name: Discover changed skill files
+        id: discover
+        env:
+          SVELTE_DEV_ROOT: ${{ runner.temp }}/svelte.dev
+        run: |
+          skill_files=$(git -C "$SVELTE_DEV_ROOT" diff --name-only --diff-filter=ACMR HEAD~1 HEAD | grep '^apps/svelte.dev/content/docs/.*\.md$' | xargs -I{} grep -l '^skill: *true' "$SVELTE_DEV_ROOT/{}" || true)
+          echo "skill_files=$skill_files" >> "$GITHUB_OUTPUT"
+
+      - name: Sync skills
+        if: steps.discover.outputs.skill_files != ''
+        env:
+          SVELTE_DEV_ROOT: ${{ runner.temp }}/svelte.dev
+          DOCS_PREFIX: apps/svelte.dev/content/docs/
+        run: |
+          for full_path in ${{ steps.discover.outputs.skill_files }}; do
+            file="${full_path#$SVELTE_DEV_ROOT/}"
+            name=$(grep '^name: ' "$full_path" | head -1 | sed 's/^name: *//')
+            repo="${file#$DOCS_PREFIX}"
+            repo="${repo#/}"
+            repo="${repo%%/*}"
+
+            output_dir="tools/skills/$name"
+            rm -rf "$output_dir"
+            mkdir -p "$output_dir"
+
+            pnpm resolve-references --file "$full_path" --repo "$repo" --output "$output_dir"
+          done
+
+      - name: Sync plugins
+        if: steps.discover.outputs.skill_files != ''
+        run: |
+          pnpm sync-claude-plugin
+          pnpm sync-cursor-plugin
+          pnpm sync-opencode-plugin
+          pnpm generate-skill-docs
+          pnpm bump-plugin-versions
+
+      - name: Check for changes
+        id: git-check
+        run: |
+          git diff --exit-code -- tools/skills/ plugins/ packages/opencode/ documentation/docs/ || echo "changed=true" >> "$GITHUB_OUTPUT"
+
+      - name: Create Pull Request
+        if: steps.git-check.outputs.changed == 'true'
+        uses: peter-evans/create-pull-request@c0f553fe549906ede9cf27b5156039d195d2ece0 # v8.1.0
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          commit-message: 'chore: sync skills from svelte.dev'
+          branch: chore/sync-skills
+          delete-branch: true
+          title: 'chore: sync skills from svelte.dev'
+          body: |
+            ## Summary
+            Automatically synced skill markdown from `sveltejs/svelte.dev` into `tools/skills/`.
+
+            ## Changes
+            - Cloned `sveltejs/svelte.dev`
+            - Filtered markdown files with `skill: true` frontmatter
+            - Rebuilt synced skill folders with `scripts/resolve-references.ts`
+            - Synced `plugins/claude/svelte/` (skills, agents)
+            - Synced `plugins/cursor/svelte/` (skills, agents, rules)
+            - Synced `packages/opencode/` (skills, instructions)
+            - Updated documentation
+
+            ## Generated by
+            GitHub Action: Sync Skills
+          labels: |
+            chore
+            automated

.github/workflows/sync-plugins.yml

@@ -0,0 +1,95 @@
+name: Sync Plugins
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - 'tools/**'
+
+permissions:
+  contents: write
+  pull-requests: write
+
+jobs:
+  sync-plugins:
+    # prevents this action from running on forks
+    if: github.repository == 'sveltejs/ai-tools'
+    name: Sync Plugins from tools/
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Setup Node.js
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: false # pnpm is not installed yet
+
+      - name: Install pnpm
+        shell: bash
+        run: |
+          PNPM_VER=$(jq -r '.packageManager | if .[0:5] == "pnpm@" then .[5:] else "packageManager in package.json does not start with pnpm@\n" | halt_error(1)  end' package.json)
+          echo installing pnpm version $PNPM_VER
+          npm i -g pnpm@$PNPM_VER
+
+      - name: Setup Node.js with pnpm cache
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: true # caches pnpm via packageManager field in package.json
+          cache: 'pnpm'
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile --prefer-offline --ignore-scripts
+
+      - name: Sync plugins
+        run: pnpm sync-plugins
+
+      - name: Generate skills documentation
+        run: pnpm generate-skill-docs
+
+      - name: Generate subagent documentation
+        run: pnpm generate-subagent-docs
+
+      - name: Check for changes
+        id: git-check
+        run: |
+          git diff --exit-code \
+            plugins/claude/svelte/ \
+            plugins/cursor/svelte/ \
+            packages/opencode/skills/ \
+            packages/opencode/instructions/ \
+            packages/opencode/schema.json \
+            documentation/docs/ \
+            || echo "changed=true" >> $GITHUB_OUTPUT
+
+      - name: Create Pull Request
+        if: steps.git-check.outputs.changed == 'true'
+        uses: peter-evans/create-pull-request@c0f553fe549906ede9cf27b5156039d195d2ece0 # v8.1.0
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          commit-message: 'chore: sync plugins from tools/'
+          branch: chore/sync-plugins
+          delete-branch: true
+          title: 'chore: sync plugins from tools/'
+          body: |
+            ## Summary
+            Automatically synced all plugins from the `tools/` source of truth.
+
+            This PR was triggered by changes to `tools/**`.
+
+            ## Changes
+            - Synced `plugins/claude/svelte/` (skills, agents with `permissionMode`)
+            - Synced `plugins/cursor/svelte/` (skills, agents, rules)
+            - Synced `packages/opencode/` (skills, instructions, schema)
+            - Updated documentation
+
+            ## Generated by
+            GitHub Action: Sync Plugins
+          labels: |
+            chore
+            automated

.github/workflows/test.yml

@@ -13,17 +13,17 @@ jobs:
 
     steps:
       - name: Checkout Repository
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
 
       - name: Setup pnpm
-        uses: pnpm/action-setup@v4
+        uses: pnpm/action-setup@v5
         with:
-          version: 10.19.0
+          version: 10.28.2
 
       - name: Setup Node.js
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
-          node-version: '22'
+          node-version: '24'
           cache: 'pnpm'
 
       - name: Install dependencies
@@ -32,13 +32,9 @@ jobs:
       - name: Build project
         run: pnpm run build
         env:
-          DATABASE_URL: file:test.db
           VOYAGE_API_KEY: dummy-key
-          DATABASE_TOKEN: dummy-key
 
       - name: Run tests
         run: pnpm run test
         env:
-          DATABASE_URL: file:test.db
           VOYAGE_API_KEY: dummy-key
-          DATABASE_TOKEN: dummy-key

.github/workflows/update-opencode-jsonschema.yml

@@ -0,0 +1,78 @@
+name: Update OpenCode JSON Schema
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - 'packages/opencode/config.ts'
+
+permissions:
+  contents: write
+  pull-requests: write
+
+jobs:
+  update-docs:
+    # prevents this action from running on forks
+    if: github.repository == 'sveltejs/ai-tools'
+    name: Update OpenCode JSON Schema
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Setup Node.js
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: false # pnpm is not installed yet
+
+      - name: Install pnpm
+        shell: bash
+        run: |
+          PNPM_VER=$(jq -r '.packageManager | if .[0:5] == "pnpm@" then .[5:] else "packageManager in package.json does not start with pnpm@\n" | halt_error(1)  end' package.json)
+          echo installing pnpm version $PNPM_VER
+          npm i -g pnpm@$PNPM_VER
+
+      - name: Setup Node.js with pnpm cache
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+          package-manager-cache: true # caches pnpm via packageManager field in package.json
+          cache: 'pnpm'
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile --prefer-offline --ignore-scripts
+
+      - name: Generate opencode JSON schema
+        run: pnpm generate-opencode-jsonschema
+
+      - name: Check for changes
+        id: git-check
+        run: |
+          git diff --exit-code packages/opencode/schema.json || echo "changed=true" >> $GITHUB_OUTPUT
+
+      - name: Create Pull Request
+        if: steps.git-check.outputs.changed == 'true'
+        uses: peter-evans/create-pull-request@c0f553fe549906ede9cf27b5156039d195d2ece0 # v8.1.0
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          commit-message: 'docs: update opencode JSON schema'
+          branch: docs/update-opencode-jsonschema
+          delete-branch: true
+          title: 'docs: update opencode JSON schema'
+          body: |
+            ## Summary
+            Automatically generated update for OpenCode JSON schema.
+
+            This PR was triggered by changes to the OpenCode configuration file `packages/opencode/config.ts`.
+
+            ## Changes
+            - Updated `packages/opencode/schema.json` with latest JSON schema
+
+            ## Generated by
+            GitHub Action: Update OpenCode JSON Schema
+          labels: |
+            automated

.github/workflows/update-prompt-docs.yml

@@ -14,17 +14,17 @@ permissions:
 jobs:
   update-docs:
     # prevents this action from running on forks
-    if: github.repository == 'sveltejs/mcp'
+    if: github.repository == 'sveltejs/ai-tools'
     name: Update Prompt Documentation
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
         with:
           fetch-depth: 0
 
       - name: Setup Node.js
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
           node-version: 24
           package-manager-cache: false # pnpm is not installed yet
@@ -37,10 +37,11 @@ jobs:
           npm i -g pnpm@$PNPM_VER
 
       - name: Setup Node.js with pnpm cache
-        uses: actions/setup-node@v6
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
         with:
           node-version: 24
           package-manager-cache: true # caches pnpm via packageManager field in package.json
+          cache: 'pnpm'
 
       - name: Install dependencies
         run: pnpm install --frozen-lockfile --prefer-offline --ignore-scripts
@@ -55,7 +56,7 @@ jobs:
 
       - name: Create Pull Request
         if: steps.git-check.outputs.changed == 'true'
-        uses: peter-evans/create-pull-request@v7
+        uses: peter-evans/create-pull-request@c0f553fe549906ede9cf27b5156039d195d2ece0 # v8.1.0
         with:
           token: ${{ secrets.GITHUB_TOKEN }}
           commit-message: 'docs: update prompts documentation'

.prettierignore

@@ -12,4 +12,6 @@ bun.lockb
 
 # Claude Code
 .claude/
-.changeset/
+.changeset/
+
+/packages/opencode/schema.json

.prettierrc

@@ -10,6 +10,12 @@
 			"options": {
 				"parser": "svelte"
 			}
+		},
+		{
+			"files": "**/references/*.md",
+			"options": {
+				"embeddedLanguageFormatting": "off"
+			}
 		}
 	]
 }

CLAUDE.md

@@ -83,7 +83,6 @@ Located in `src/lib/server/analyze/`:
 
 Required environment variables:
 
-- `DATABASE_URL`: SQLite database path (default: `file:test.db`)
 - `VOYAGE_API_KEY`: API key for embeddings support (optional)
 
 When connected to the svelte-llm MCP server, you have access to comprehensive Svelte 5 and SvelteKit documentation. Here's how to use the available tools effectively:

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 [Svelte Contributors](https://github.com/sveltejs/ai-tools/graphs/contributors)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

apps/mcp-remote/.env.example

@@ -1,3 +1 @@
-DATABASE_URL=file:test.db
-DATABASE_TOKEN=needs_to_be_set_but_it_can_be_anything
 VOYAGE_API_KEY=your_actual_api_key_here

apps/mcp-remote/drizzle.config.ts

@@ -1,12 +0,0 @@
-import { defineConfig } from 'drizzle-kit';
-
-if (!process.env.DATABASE_URL) throw new Error('DATABASE_URL is not set');
-if (!process.env.DATABASE_TOKEN) throw new Error('DATABASE_TOKEN is not set');
-
-export default defineConfig({
-	schema: './src/lib/server/db/schema.ts',
-	dialect: 'turso',
-	dbCredentials: { url: process.env.DATABASE_URL, authToken: process.env.DATABASE_TOKEN },
-	verbose: true,
-	strict: true,
-});

apps/mcp-remote/package.json

@@ -23,10 +23,6 @@
 		"test:unit": "vitest",
 		"test": "npm run test:unit -- --run",
 		"test:watch": "npm run test:unit -- --watch",
-		"db:push": "drizzle-kit push",
-		"db:generate": "drizzle-kit generate",
-		"db:migrate": "drizzle-kit migrate",
-		"db:studio": "drizzle-kit studio",
 		"inspect": "pnpm mcp-inspector"
 	},
 	"keywords": [
@@ -37,34 +33,31 @@
 	],
 	"private": true,
 	"devDependencies": {
-		"@eslint/compat": "^1.3.2",
-		"@eslint/js": "^9.36.0",
-		"@libsql/client": "^0.15.0",
-		"@modelcontextprotocol/inspector": "^0.17.0",
-		"@sveltejs/adapter-vercel": "^6.0.0",
-		"@sveltejs/kit": "^2.22.0",
-		"@sveltejs/vite-plugin-svelte": "^6.0.0",
-		"@types/node": "^24.3.1",
-		"@typescript-eslint/parser": "^8.44.0",
-		"drizzle-kit": "^0.31.0",
-		"drizzle-orm": "^0.44.0",
-		"eslint-config-prettier": "^10.0.1",
-		"eslint-plugin-svelte": "^3.12.3",
-		"globals": "^16.0.0",
-		"prettier": "^3.4.2",
-		"prettier-plugin-svelte": "^3.3.3",
-		"svelte": "^5.0.0",
-		"svelte-check": "^4.0.0",
-		"svelte-eslint-parser": "^1.3.2",
-		"typescript": "^5.0.0",
-		"vite": "^7.0.4",
-		"vite-plugin-devtools-json": "^1.0.0",
-		"vitest": "^4.0.0"
+		"@eslint/compat": "catalog:lint",
+		"@eslint/js": "catalog:lint",
+		"@modelcontextprotocol/inspector": "catalog:ai",
+		"@sveltejs/adapter-vercel": "catalog:svelte",
+		"@sveltejs/kit": "catalog:svelte",
+		"@sveltejs/vite-plugin-svelte": "catalog:svelte",
+		"@types/node": "catalog:tooling",
+		"@typescript-eslint/parser": "catalog:lint",
+		"eslint-config-prettier": "catalog:lint",
+		"eslint-plugin-svelte": "catalog:lint",
+		"globals": "catalog:lint",
+		"prettier": "catalog:lint",
+		"prettier-plugin-svelte": "catalog:lint",
+		"svelte": "catalog:svelte",
+		"svelte-check": "catalog:svelte",
+		"svelte-eslint-parser": "catalog:lint",
+		"typescript": "catalog:tooling",
+		"vite": "catalog:tooling",
+		"vite-plugin-devtools-json": "catalog:tooling",
+		"vitest": "catalog:tooling"
 	},
 	"dependencies": {
-		"@sveltejs/mcp-schema": "workspace:^",
 		"@sveltejs/mcp-server": "workspace:^",
-		"@tmcp/transport-http": "^0.7.1",
-		"tmcp": "^1.15.3"
+		"@tmcp/transport-http": "catalog:tmcp",
+		"@vercel/analytics": "catalog:tooling",
+		"tmcp": "catalog:tmcp"
 	}
 }

apps/mcp-remote/src/hooks.server.ts

@@ -1,11 +1,9 @@
+import { dev } from '$app/environment';
 import { http_transport } from '$lib/mcp/index.js';
-import { db } from '$lib/server/db/index.js';
 import { redirect } from '@sveltejs/kit';
+import { track } from '@vercel/analytics/server';
 
 export async function handle({ event, resolve }) {
-	if (event.request.method === 'POST') {
-		console.log(await event.request.clone().text());
-	}
 	if (event.request.method === 'GET') {
 		const accept = event.request.headers.get('accept');
 		if (accept) {
@@ -18,24 +16,12 @@ export async function handle({ event, resolve }) {
 		}
 	}
 	const mcp_response = await http_transport.respond(event.request, {
-		db,
+		// only add analytics in production
+		track: dev
+			? undefined
+			: async (session_id, event, extra) => {
+					await track(event, { session_id, ...(extra ? { extra } : {}) });
+				},
 	});
-	// we are deploying on vercel the SSE connection will timeout after 5 minutes...for
-	// the moment we are not sending back any notifications (logs, or list changed notifications)
-	// so it's a waste of resources to keep a connection open that will error
-	// after 5 minutes making the logs dirty. For this reason if we have a response from
-	// the MCP server and it's a GET request we just return an empty response (it has to be
-	// 200 or the MCP client will complain)
-	if (mcp_response && event.request.method === 'GET') {
-		try {
-			return mcp_response;
-		} finally {
-			try {
-				await mcp_response.body?.cancel();
-			} catch {
-				// ignore
-			}
-		}
-	}
 	return mcp_response ?? resolve(event);
 }

apps/mcp-remote/src/lib/mcp/index.ts

@@ -4,4 +4,9 @@ import { HttpTransport } from '@tmcp/transport-http';
 export const http_transport = new HttpTransport(server, {
 	cors: true,
 	path: '/mcp',
+	// we are deploying on vercel the SSE connection will timeout after 5 minutes...for
+	// the moment we are not sending back any notifications (logs, or list changed notifications)
+	// so it's a waste of resources to keep a connection open that will error
+	// after 5 minutes making the logs dirty.
+	disableSse: true,
 });

apps/mcp-remote/src/lib/server/db/index.ts

@@ -1,15 +0,0 @@
-import { createClient } from '@libsql/client';
-import { drizzle } from 'drizzle-orm/libsql';
-import * as schema from './schema.js';
-// let's disable it for the moment...i can't figure out a way to make it wotk with eslint
-// eslint-disable-next-line import/extensions
-import { DATABASE_TOKEN, DATABASE_URL } from '$env/static/private';
-if (!DATABASE_URL) throw new Error('DATABASE_URL is not set');
-if (!DATABASE_TOKEN) throw new Error('DATABASE_TOKEN is not set');
-
-const client = createClient({
-	url: DATABASE_URL,
-	authToken: DATABASE_TOKEN,
-});
-
-export const db = drizzle(client, { schema, logger: true });

apps/mcp-remote/src/lib/server/db/schema.ts

@@ -1,2 +0,0 @@
-// we need to re-export from here to allow for the drizzle config to pick them up for migrations
-export * from '@sveltejs/mcp-schema/schema';

documentation/docs/10-introduction/10-overview.md

@@ -2,43 +2,11 @@
 title: Overview
 ---
 
-The Svelte MCP ([Model Context Protocol](https://modelcontextprotocol.io/docs/getting-started/intro)) server can help your LLM or agent of choice write better Svelte code. It works by providing documentation relevant to the task at hand, and statically analysing generated code so that it can suggest fixes and best practices.
+The following pages will help you set up and use the AI tools officially maintained by the Svelte team.
 
-## Setup
+There are four tools, designed to help your agent write correct, robust Svelte code. They are designed to work together, but each can be used individually:
 
-The setup varies based on the version of the MCP you prefer — remote or local — and your chosen MCP client (e.g. Claude Code, Codex CLI or GitHub Copilot):
-
-- [local setup](local-setup) using `@sveltejs/mcp`
-- [remote setup](remote-setup) using `https://mcp.svelte.dev/mcp`
-
-## Usage
-
-To get the most out of the MCP server we recommend including the following prompt in your [`AGENTS.md`](https://agents.md) (or [`CLAUDE.md`](https://docs.claude.com/en/docs/claude-code/memory#claude-md-imports), if using Claude Code). This will tell the LLM which tools are available and when it's appropriate to use them.
-
-```md
-You are able to use the Svelte MCP server, where you have access to comprehensive Svelte 5 and SvelteKit documentation. Here's how to use the available tools effectively:
-
-## Available MCP Tools:
-
-### 1. list-sections
-
-Use this FIRST to discover all available documentation sections. Returns a structured list with titles, use_cases, and paths.
-When asked about Svelte or SvelteKit topics, ALWAYS use this tool at the start of the chat to find relevant sections.
-
-### 2. get-documentation
-
-Retrieves full documentation content for specific sections. Accepts single or multiple sections.
-After calling the list-sections tool, you MUST analyze the returned documentation sections (especially the use_cases field) and then use the get-documentation tool to fetch ALL documentation sections that are relevant for the user's task.
-
-### 3. svelte-autofixer
-
-Analyzes Svelte code and returns issues and suggestions.
-You MUST use this tool whenever writing Svelte code before sending it to the user. Keep calling it until no issues or suggestions are returned.
-
-### 4. playground-link
-
-Generates a Svelte Playground link with the provided code.
-After completing the code, ask the user if they want a playground link. Only call this tool after user confirmation and NEVER if code was written to files in their project.
-```
-
-If your MCP client supports it, we also recommend using the [svelte-task](prompts#svelte-task) prompt to instruct the LLM on the best way to use the MCP server.
+- [Instructions](instructions): small prompt always injected into your session to make your agent more aware of the available tools
+- [MCP Server](mcp): with tools, prompts and resources to give your agent more context, by pulling directly from the official Svelte documentation and using static analysis to correct common generative AI pitfalls
+- [Skills](skills): lazy-loaded descriptions that teach your agent Svelte best practices, and how to use the [`@sveltejs/mcp` cli](cli)
+- [Subagents](subagent): focused agents that can be invoked in parallel to execute atomic operations in a separate context window

documentation/docs/20-instructions/.generated/agents.md

@@ -0,0 +1,23 @@
+You are able to use the Svelte MCP server, where you have access to comprehensive Svelte 5 and SvelteKit documentation. Here's how to use the available tools effectively:
+
+## Available Svelte MCP Tools:
+
+### 1. list-sections
+
+Use this FIRST to discover all available documentation sections. Returns a structured list with titles, use_cases, and paths.
+When asked about Svelte or SvelteKit topics, ALWAYS use this tool at the start of the chat to find relevant sections.
+
+### 2. get-documentation
+
+Retrieves full documentation content for specific sections. Accepts single or multiple sections.
+After calling the list-sections tool, you MUST analyze the returned documentation sections (especially the use_cases field) and then use the get-documentation tool to fetch ALL documentation sections that are relevant for the user's task.
+
+### 3. svelte-autofixer
+
+Analyzes Svelte code and returns issues and suggestions.
+You MUST use this tool whenever writing Svelte code before sending it to the user. Keep calling it until no issues or suggestions are returned.
+
+### 4. playground-link
+
+Generates a Svelte Playground link with the provided code.
+After completing the code, ask the user if they want a playground link. Only call this tool after user confirmation and NEVER if code was written to files in their project.

documentation/docs/20-instructions/10-instructions.md

@@ -0,0 +1,13 @@
+---
+title: AGENTS.md
+---
+
+To get the most out of the [MCP server](mcp) and [skills](skills) we recommend including the following prompt in your [`AGENTS.md`](https://agents.md) (or [`CLAUDE.md`](https://docs.claude.com/en/docs/claude-code/memory#claude-md-imports) or [`GEMINI.md`](https://geminicli.com/docs/cli/gemini-md/), if using Claude Code or Gemini). This will tell your agent which tools are available and when it is appropriate to use them.
+
+> [!NOTE] This is already setup for you when using `npx sv add mcp`
+
+<!-- prettier-ignore-start -->
+````markdown
+@include .generated/agents.md
+````
+<!-- prettier-ignore-end -->

documentation/docs/20-instructions/index.md

@@ -0,0 +1,3 @@
+---
+title: Instructions
+---

documentation/docs/20-setup/index.md

@@ -1,3 +0,0 @@
----
-title: Setup
----

documentation/docs/30-capabilities/index.md

@@ -1,3 +0,0 @@
----
-title: Capabilities
----

documentation/docs/30-mcp/.generated/prompts.md

@@ -1,9 +1,3 @@
----
-title: Prompts
----
-
-This is the list of available prompts provided by the MCP server. Prompts are selected by the user and are sent as a user message. They can be useful to write repetitive instructions for the LLM on how to properly use the MCP server.
-
 ## svelte-task
 
 This prompt should be used whenever you are asking the model to work on a Svelte-related task. It will instruct the LLM which documentation sections are available, which tools to invoke, when to invoke them, and how to interpret the results.
@@ -11,10 +5,22 @@ This prompt should be used whenever you are asking the model to work on a Svelte
 <details>
 	<summary>Copy the prompt</summary>
 
-```md
-You are a Svelte expert tasked to build components and utilities for Svelte developers. If you need documentation for anything related to Svelte you can invoke the tool `get_documentation` with one of the following paths:
+<!-- prettier-ignore-start -->
+````markdown
+You are a Svelte expert tasked to build components and utilities for Svelte developers. If you need documentation for anything related to Svelte you can invoke the tool `get-documentation` with one of the following paths. However: before invoking the `get-documentation` tool, try to answer the users query using your own knowledge and the `svelte-autofixer` tool. Be mindful of how many section you request, since it is token-intensive!
 <available-docs>
 
+- title: Overview, use_cases: use title and path to estimate use case, path: ai/overview
+- title: Local setup, use_cases: use title and path to estimate use case, path: ai/local-setup
+- title: Remote setup, use_cases: use title and path to estimate use case, path: ai/remote-setup
+- title: Tools, use_cases: use title and path to estimate use case, path: ai/tools
+- title: Resources, use_cases: use title and path to estimate use case, path: ai/resources
+- title: Prompts, use_cases: use title and path to estimate use case, path: ai/prompts
+- title: Overview, use_cases: use title and path to estimate use case, path: ai/plugin
+- title: Subagent, use_cases: use title and path to estimate use case, path: ai/subagent
+- title: Overview, use_cases: use title and path to estimate use case, path: ai/opencode-plugin
+- title: Subagent, use_cases: use title and path to estimate use case, path: ai/opencode-subagent
+- title: Overview, use_cases: use title and path to estimate use case, path: ai/skills
 - title: Overview, use_cases: project setup, creating new svelte apps, scaffolding, cli tools, initializing projects, path: cli/overview
 - title: Frequently asked questions, use_cases: project setup, initializing new svelte projects, troubleshooting cli installation, package manager configuration, path: cli/faq
 - title: sv create, use_cases: project setup, starting new sveltekit app, initializing project, creating from playground, choosing project template, path: cli/sv-create
@@ -24,7 +30,8 @@ You are a Svelte expert tasked to build components and utilities for Svelte deve
 - title: devtools-json, use_cases: development setup, chrome devtools integration, browser-based editing, local development workflow, debugging setup, path: cli/devtools-json
 - title: drizzle, use_cases: database setup, sql queries, orm integration, data modeling, postgresql, mysql, sqlite, server-side data access, database migrations, type-safe queries, path: cli/drizzle
 - title: eslint, use_cases: code quality, linting, error detection, project setup, code standards, team collaboration, typescript projects, path: cli/eslint
-- title: lucia, use_cases: authentication, login systems, user management, registration pages, session handling, auth setup, path: cli/lucia
+- title: better-auth, use_cases: use title and path to estimate use case, path: cli/better-auth
+- title: mcp, use_cases: use title and path to estimate use case, path: cli/mcp
 - title: mdsvex, use_cases: blog, content sites, markdown rendering, documentation sites, technical writing, cms integration, article pages, path: cli/mdsvex
 - title: paraglide, use_cases: internationalization, multi-language sites, i18n, translation, localization, language switching, global apps, multilingual content, path: cli/paraglide
 - title: playwright, use_cases: browser testing, e2e testing, integration testing, test automation, quality assurance, ci/cd pipelines, testing user flows, path: cli/playwright
@@ -33,6 +40,8 @@ You are a Svelte expert tasked to build components and utilities for Svelte deve
 - title: sveltekit-adapter, use_cases: deployment, production builds, hosting setup, choosing deployment platform, configuring adapters, static site generation, node server, vercel, cloudflare, netlify, path: cli/sveltekit-adapter
 - title: tailwindcss, use_cases: project setup, styling, css framework, rapid prototyping, utility-first css, design systems, responsive design, adding tailwind to svelte, path: cli/tailwind
 - title: vitest, use_cases: testing, unit tests, component testing, test setup, quality assurance, ci/cd pipelines, test-driven development, path: cli/vitest
+- title: add-on, use_cases: use title and path to estimate use case, path: cli/add-on
+- title: sv-utils, use_cases: use title and path to estimate use case, path: cli/sv-utils
 - title: Introduction, use_cases: learning sveltekit, project setup, understanding framework basics, choosing between svelte and sveltekit, getting started with full-stack apps, path: kit/introduction
 - title: Creating a project, use_cases: project setup, starting new sveltekit app, initial development environment, first-time sveltekit users, scaffolding projects, path: kit/creating-a-project
 - title: Project types, use_cases: deployment, project setup, choosing adapters, ssg, spa, ssr, serverless, mobile apps, desktop apps, pwa, offline apps, browser extensions, separate backend, docker containers, path: kit/project-types
@@ -100,12 +109,6 @@ You are a Svelte expert tasked to build components and utilities for Svelte deve
 - title: Configuration, use_cases: project setup, configuration, adapters, deployment, build settings, environment variables, routing customization, prerendering, csp security, csrf protection, path configuration, typescript setup, path: kit/configuration
 - title: Command Line Interface, use_cases: project setup, typescript configuration, generated types, ./$types imports, initial project configuration, path: kit/cli
 - title: Types, use_cases: typescript, type safety, route parameters, api endpoints, load functions, form actions, generated types, jsconfig setup, path: kit/types
-- title: Overview, use_cases: use title and path to estimate use case, path: mcp/overview
-- title: Local setup, use_cases: use title and path to estimate use case, path: mcp/local-setup
-- title: Remote setup, use_cases: use title and path to estimate use case, path: mcp/remote-setup
-- title: Tools, use_cases: use title and path to estimate use case, path: mcp/tools
-- title: Resources, use_cases: use title and path to estimate use case, path: mcp/resources
-- title: Prompts, use_cases: use title and path to estimate use case, path: mcp/prompts
 - title: Overview, use_cases: always, any svelte project, getting started, learning svelte, introduction, project setup, understanding framework basics, path: svelte/overview
 - title: Getting started, use_cases: project setup, starting new svelte project, initial installation, choosing between sveltekit and vite, editor configuration, path: svelte/getting-started
 - title: .svelte files, use_cases: always, any svelte project, component creation, project setup, learning svelte basics, path: svelte/svelte-files
@@ -152,6 +155,8 @@ You are a Svelte expert tasked to build components and utilities for Svelte deve
 - title: Context, use_cases: shared state, avoiding prop drilling, component communication, theme providers, user context, authentication state, configuration sharing, deeply nested components, path: svelte/context
 - title: Lifecycle hooks, use_cases: component initialization, cleanup tasks, timers, subscriptions, dom measurements, chat windows, autoscroll features, migration from svelte 4, path: svelte/lifecycle-hooks
 - title: Imperative component API, use_cases: project setup, client-side rendering, server-side rendering, ssr, hydration, testing, programmatic component creation, tooltips, dynamic mounting, path: svelte/imperative-component-api
+- title: Hydratable data, use_cases: use title and path to estimate use case, path: svelte/hydratable
+- title: Best practices, use_cases: use title and path to estimate use case, path: svelte/best-practices
 - title: Testing, use_cases: testing, quality assurance, unit tests, integration tests, component tests, e2e tests, vitest setup, playwright setup, test automation, path: svelte/testing
 - title: TypeScript, use_cases: typescript setup, type safety, component props typing, generic components, wrapper components, dom type augmentation, project configuration, path: svelte/typescript
 - title: Custom elements, use_cases: web components, custom elements, component library, design system, framework-agnostic components, embedding svelte in non-svelte apps, shadow dom, path: svelte/custom-elements
@@ -191,6 +196,8 @@ You are a Svelte expert tasked to build components and utilities for Svelte deve
 
 </available-docs>
 
+These are the available documentation sections that `list-sections` will return, you do not need to call it again.
+
 Every time you write a Svelte component or a Svelte module you MUST invoke the `svelte-autofixer` tool providing the code. The tool will return a list of issues or suggestions. If there are any issues or suggestions you MUST fix them and call the tool again with the updated code. You MUST keep doing this until the tool returns no issues or suggestions. Only then you can return the code to the user.
 
 This is the task you will work on:
@@ -200,6 +207,7 @@ This is the task you will work on:
 </task>
 
 If you are not writing the code into a file, once you have the final version of the code ask the user if it wants to generate a playground link to quickly check the code in it and if it answer yes call the `playground-link` tool and return the url to the user nicely formatted. The playground link MUST be generated only once you have the final version of the code and you are ready to share it, it MUST include an entry point file called `App.svelte` where the main component should live. If you have multiple files to include in the playground link you can include them all at the root.
-```
+````
+<!-- prettier-ignore-end -->
 
 </details>

documentation/docs/30-mcp/10-mcp.md

@@ -0,0 +1,16 @@
+---
+title: Overview
+---
+
+The Svelte MCP ([Model Context Protocol](https://modelcontextprotocol.io/docs/getting-started/intro)) server can help your agent write better Svelte code. It works by providing relevant documentation, and statically analysing generated code so that it can suggest fixes and best practices.
+
+## Setup
+
+The setup varies based on the version of the MCP you prefer — remote or local — and your chosen MCP client (e.g. Claude Code, Codex CLI or GitHub Copilot):
+
+- [local setup](local-setup) using `@sveltejs/mcp`
+- [remote setup](remote-setup) using `https://mcp.svelte.dev/mcp`
+
+## Usage
+
+If your MCP client supports it, we also recommend using the [svelte-task](prompts#svelte-task) prompt to instruct the LLM on the best way to use the MCP server.

documentation/docs/30-mcp/20-local-setup.md

@@ -45,6 +45,27 @@ command = "npx"
 args = ["-y", "@sveltejs/mcp"]
 ```
 
+## Copilot CLI
+
+Use the Copilot CLI to interactively add the MCP server:
+
+```bash
+/mcp add
+```
+
+Alternatively, create or edit `~/.copilot/mcp-config.json` and add the following configuration:
+
+```json
+{
+	"mcpServers": {
+		"svelte": {
+			"command": "npx",
+			"args": ["-y", "@sveltejs/mcp"]
+		}
+	}
+}
+```
+
 ## Gemini CLI
 
 To include the local MCP version in Gemini CLI, simply run the following command:
@@ -57,7 +78,7 @@ The `[scope]` must be `user`, `project` or `local`.
 
 ## OpenCode
 
-Run the command:
+You can automatically configure the MCP server using the [OpenCode plugin](opencode-plugin) (recommended). If you prefer to configure the MCP server manually, run:
 
 ```bash
 opencode mcp add
@@ -91,6 +112,8 @@ opencode mcp add
 
 ## Cursor
 
+You can automatically configure the MCP server using the [Cursor plugin](cursor-plugin) (recommended). If you prefer to configure the MCP server manually you can:
+
 - Open the command palette
 - Select "View: Open MCP Settings"
 - Click on "Add custom MCP"
@@ -110,6 +133,12 @@ It will open a file with your MCP servers where you can add the following config
 
 ## Zed
 
+Install the [Svelte MCP Server extension](https://zed.dev/extensions/svelte-mcp).
+
+<details>
+
+<summary>Configure Manually</summary>
+
 - Open the command palette
 - Search and select "agent:open settings"
 - In settings panel look for `Model Context Protocol (MCP) Servers`
@@ -127,6 +156,8 @@ It will open a popup with MCP server config where you can add the following conf
 }
 ```
 
+</details>
+
 ## Other clients
 
 If we didn't include the MCP client you are using, refer to their documentation for `stdio` servers and use `npx` as the command and `-y @sveltejs/mcp` as the arguments.

documentation/docs/30-mcp/30-remote-setup.md

@@ -16,6 +16,8 @@ claude mcp add -t http -s [scope] svelte https://mcp.svelte.dev/mcp
 
 You can choose your preferred `scope` (it must be `user`, `project` or `local`) and `name`.
 
+If you prefer you can also install the `svelte` plugin in [the Svelte Claude Code Marketplace](claude-plugin) that will give you both the remote server and useful [skills](skills).
+
 ## Claude Desktop
 
 - Open Settings > Connectors
@@ -34,6 +36,26 @@ experimental_use_rmcp_client = true
 url = "https://mcp.svelte.dev/mcp"
 ```
 
+## Copilot CLI
+
+Use the Copilot CLI to interactively add the MCP server:
+
+```bash
+/mcp add
+```
+
+Alternatively, create or edit `~/.copilot/mcp-config.json` and add the following configuration:
+
+```json
+{
+	"mcpServers": {
+		"svelte": {
+			"url": "https://mcp.svelte.dev/mcp"
+		}
+	}
+}
+```
+
 ## Gemini CLI
 
 To use the remote MCP server with Gemini CLI, simply run the following command:
@@ -42,11 +64,11 @@ To use the remote MCP server with Gemini CLI, simply run the following command:
 gemini mcp add -t http -s [scope] svelte https://mcp.svelte.dev/mcp
 ```
 
-The `[scope]` must be `user`, `project` or `local`.
+The `[scope]` must be `user` or `project`.
 
 ## OpenCode
 
-Run the command:
+You can automatically configure the MCP server using the [OpenCode plugin](opencode-plugin) (recommended). If you prefer to configure the MCP server manually, run:
 
 ```bash
 opencode mcp add
@@ -80,6 +102,8 @@ opencode mcp add
 
 ## Cursor
 
+You can automatically configure the MCP server using the [Cursor plugin](cursor-plugin) (recommended). If you prefer to configure the MCP server manually you can:
+
 - Open the command palette
 - Select "View: Open MCP Settings"
 - Click on "Add custom MCP"

documentation/docs/30-mcp/60-prompts.md

@@ -0,0 +1,7 @@
+---
+title: Prompts
+---
+
+This is the list of available prompts provided by the MCP server. Prompts are selected by the user and are sent as a user message. They can be useful to write repetitive instructions for the LLM on how to properly use the MCP server.
+
+@include .generated/prompts.md

documentation/docs/30-mcp/70-cli.md

@@ -0,0 +1,78 @@
+---
+title: CLI
+---
+
+The `@sveltejs/mcp` npm package normally launches the local `stdio` MCP server:
+
+```bash
+npx -y @sveltejs/mcp
+```
+
+If you invoke it with a subcommand, it behaves like a regular CLI and prints the result directly in your terminal instead. This is useful for agents, scripts and quick manual checks.
+
+## Usage
+
+```bash
+npx -y @sveltejs/mcp <command> [options]
+```
+
+Available commands:
+
+- `list-sections`
+- `get-documentation <sections>`
+- `svelte-autofixer <code_or_path>`
+
+You can learn more about the commands with
+
+```bash
+npx -y @sveltejs/mcp --help
+npx -y @sveltejs/mcp <command> --help
+npx -y @sveltejs/mcp --version
+```
+
+## `list-sections`
+
+Lists all available Svelte and SvelteKit documentation sections.
+
+```bash
+npx -y @sveltejs/mcp list-sections
+```
+
+The output is a structured text list of sections, including each section's title, `use_cases`, and documentation path. This is the same catalog the MCP tool uses before calling `get-documentation`.
+
+## `get-documentation`
+
+Fetches the full documentation for one or more sections.
+
+```bash
+npx -y @sveltejs/mcp get-documentation 'svelte/$state'
+# or
+npx -y @sveltejs/mcp get-documentation 'svelte/$state,svelte/await-expressions'
+```
+
+Each section can be matched by title or by documentation path. If a section cannot be found, the CLI returns an error plus similar matches when available.
+
+## `svelte-autofixer`
+
+Runs the Svelte autofixer against either inline code or a file path:
+
+```bash
+npx -y @sveltejs/mcp svelte-autofixer 'src/routes/+page.svelte'
+```
+
+If the argument is an existing path, the CLI reads the file automatically. Otherwise it treats the argument as raw Svelte code.
+
+Because most shells expand `$`, inline code should be quoted or escaped correctly. In practice, passing a file path is usually easier than passing source directly.
+
+Available options:
+
+- `--svelte-version <4|5>` - choose which Svelte version to validate against (defaults to `5`)
+- `--async` - enable async Svelte analysis for Svelte 5 projects
+
+The command prints an object with:
+
+- `issues`
+- `suggestions`
+- `require_another_tool_call_after_fixing`
+
+This makes it easy to use in an agentic loop: run the autofixer, apply fixes, then run it again until it reports no remaining issues or suggestions.

documentation/docs/30-mcp/index.md

@@ -0,0 +1,3 @@
+---
+title: MCP server
+---

documentation/docs/40-skills/.generated/skills.md

@@ -0,0 +1,263 @@
+## `svelte-code-writer`
+
+CLI tools for Svelte 5 documentation lookup and code analysis. MUST be used whenever creating, editing or analyzing any Svelte component (.svelte) or Svelte module (.svelte.ts/.svelte.js). If possible, this skill should be executed within the svelte-file-editor agent for optimal results.
+
+<a href="https://github.com/sveltejs/ai-tools/releases?q=svelte-code-writer" target="_blank" rel="noopener noreferrer">Open Releases page</a>
+
+<details>
+	<summary>View skill content</summary>
+
+<!-- prettier-ignore-start -->
+````markdown
+# Svelte 5 Code Writer
+
+## CLI Tools
+
+You have access to `@sveltejs/mcp` CLI for Svelte-specific assistance. Use these commands via `npx`:
+
+### List Documentation Sections
+
+```bash
+npx @sveltejs/mcp list-sections
+```
+
+Lists all available Svelte 5 and SvelteKit documentation sections with titles and paths.
+
+### Get Documentation
+
+```bash
+npx @sveltejs/mcp get-documentation "<section1>,<section2>,..."
+```
+
+Retrieves full documentation for specified sections. Use after `list-sections` to fetch relevant docs.
+
+**Example:**
+
+```bash
+npx @sveltejs/mcp get-documentation "$state,$derived,$effect"
+```
+
+### Svelte Autofixer
+
+```bash
+npx @sveltejs/mcp svelte-autofixer "<code_or_path>" [options]
+```
+
+Analyzes Svelte code and suggests fixes for common issues.
+
+**Options:**
+
+- `--async` - Enable async Svelte mode (default: false)
+- `--svelte-version` - Target version: 4 or 5 (default: 5)
+
+**Examples:**
+
+```bash
+# Analyze inline code (escape $ as \$)
+npx @sveltejs/mcp svelte-autofixer '<script>let count = \$state(0);</script>'
+
+# Analyze a file
+npx @sveltejs/mcp svelte-autofixer ./src/lib/Component.svelte
+
+# Target Svelte 4
+npx @sveltejs/mcp svelte-autofixer ./Component.svelte --svelte-version 4
+```
+
+**Important:** When passing code with runes (`$state`, `$derived`, etc.) via the terminal, escape the `$` character as `\$` to prevent shell variable substitution.
+
+## Workflow
+
+1. **Uncertain about syntax?** Run `list-sections` then `get-documentation` for relevant topics
+2. **Reviewing/debugging?** Run `svelte-autofixer` on the code to detect issues
+3. **Always validate** - Run `svelte-autofixer` before finalizing any Svelte component
+````
+<!-- prettier-ignore-end -->
+
+</details>
+
+## `svelte-core-bestpractices`
+
+Guidance on writing fast, robust, modern Svelte code. Load this skill whenever in a Svelte project and asked to write/edit or analyze a Svelte component or module. Covers reactivity, event handling, styling, integration with libraries and more.
+
+<a href="https://github.com/sveltejs/ai-tools/releases?q=svelte-core-bestpractices" target="_blank" rel="noopener noreferrer">Open Releases page</a>
+
+<details>
+	<summary>View skill content</summary>
+
+<!-- prettier-ignore-start -->
+````markdown
+## `$state`
+
+Only use the `$state` rune for variables that should be _reactive_ — in other words, variables that cause an `$effect`, `$derived` or template expression to update. Everything else can be a normal variable.
+
+Objects and arrays (`$state({...})` or `$state([...])`) are made deeply reactive, meaning mutation will trigger updates. This has a trade-off: in exchange for fine-grained reactivity, the objects must be proxied, which has performance overhead. In cases where you're dealing with large objects that are only ever reassigned (rather than mutated), use `$state.raw` instead. This is often the case with API responses, for example.
+
+## `$derived`
+
+To compute something from state, use `$derived` rather than `$effect`:
+
+```js
+// do this
+let square = $derived(num * num);
+
+// don't do this
+let square;
+
+$effect(() => {
+	square = num * num;
+});
+```
+
+> [!NOTE] `$derived` is given an expression, _not_ a function. If you need to use a function (because the expression is complex, for example) use `$derived.by`.
+
+Deriveds are writable — you can assign to them, just like `$state`, except that they will re-evaluate when their expression changes.
+
+If the derived expression is an object or array, it will be returned as-is — it is _not_ made deeply reactive. You can, however, use `$state` inside `$derived.by` in the rare cases that you need this.
+
+## `$effect`
+
+Effects are an escape hatch and should mostly be avoided. In particular, avoid updating state inside effects.
+
+- If you need to sync state to an external library such as D3, it is often neater to use [`{@attach ...}`](references/@attach.md)
+- If you need to run some code in response to user interaction, put the code directly in an event handler or use a [function binding](references/bind.md) as appropriate
+- If you need to log values for debugging purposes, use [`$inspect`](references/$inspect.md)
+- If you need to observe something external to Svelte, use [`createSubscriber`](references/svelte-reactivity.md)
+
+Never wrap the contents of an effect in `if (browser) {...}` or similar — effects do not run on the server.
+
+## `$props`
+
+Treat props as though they will change. For example, values that depend on props should usually use `$derived`:
+
+```js
+// @errors: 2451
+let { type } = $props();
+
+// do this
+let color = $derived(type === 'danger' ? 'red' : 'green');
+
+// don't do this — `color` will not update if `type` changes
+let color = type === 'danger' ? 'red' : 'green';
+```
+
+## `$inspect.trace`
+
+`$inspect.trace` is a debugging tool for reactivity. If something is not updating properly or running more than it should you can add `$inspect.trace(label)` as the first line of an `$effect` or `$derived.by` (or any function they call) to trace their dependencies and discover which one triggered an update.
+
+## Events
+
+Any element attribute starting with `on` is treated as an event listener:
+
+```svelte
+<button onclick={() => {...}}>click me</button>
+
+<!-- attribute shorthand also works -->
+<button {onclick}>...</button>
+
+<!-- so do spread attributes -->
+<button {...props}>...</button>
+```
+
+If you need to attach listeners to `window` or `document` you can use `<svelte:window>` and `<svelte:document>`:
+
+```svelte
+<svelte:window onkeydown={...} />
+<svelte:document onvisibilitychange={...} />
+```
+
+Avoid using `onMount` or `$effect` for this.
+
+## Snippets
+
+[Snippets](references/snippet.md) are a way to define reusable chunks of markup that can be instantiated with the [`{@render ...}`](references/@render.md) tag, or passed to components as props. They must be declared within the template.
+
+```svelte
+{#snippet greeting(name)}
+	<p>hello {name}!</p>
+{/snippet}
+
+{@render greeting('world')}
+```
+
+> [!NOTE] Snippets declared at the top level of a component (i.e. not inside elements or blocks) can be referenced inside `<script>`. A snippet that doesn't reference component state is also available in a `<script module>`, in which case it can be exported for use by other components.
+
+## Each blocks
+
+Prefer to use [keyed each blocks](references/each.md) — this improves performance by allowing Svelte to surgically insert or remove items rather than updating the DOM belonging to existing items.
+
+> [!NOTE] The key _must_ uniquely identify the object. Do not use the index as a key.
+
+Avoid destructuring if you need to mutate the item (with something like `bind:value={item.count}`, for example).
+
+## Using JavaScript variables in CSS
+
+If you have a JS variable that you want to use inside CSS you can set a CSS custom property with the `style:` directive.
+
+```svelte
+<div style:--columns={columns}>...</div>
+```
+
+You can then reference `var(--columns)` inside the component's `<style>`.
+
+## Styling child components
+
+The CSS in a component's `<style>` is scoped to that component. If a parent component needs to control the child's styles, the preferred way is to use CSS custom properties:
+
+```svelte
+<!-- Parent.svelte -->
+<Child --color="red" />
+
+<!-- Child.svelte -->
+<h1>Hello</h1>
+
+<style>
+	h1 {
+		color: var(--color);
+	}
+</style>
+```
+
+If this is impossible (for example, the child component comes from a library) you can use `:global` to override styles:
+
+```svelte
+<div>
+	<Child />
+</div>
+
+<style>
+	div :global {
+		h1 {
+			color: red;
+		}
+	}
+</style>
+```
+
+## Context
+
+Consider using context instead of declaring state in a shared module. This will scope the state to the part of the app that needs it, and eliminate the possibility of it leaking between users when server-side rendering.
+
+Use `createContext` rather than `setContext` and `getContext`, as it provides type safety.
+
+## Async Svelte
+
+If using version 5.36 or higher, you can use [await expressions](references/await-expressions.md) and [hydratable](references/hydratable.md) to use promises directly inside components. Note that these require the `experimental.async` option to be enabled in `svelte.config.js` as they are not yet considered fully stable.
+
+## Avoid legacy features
+
+Always use runes mode for new code, and avoid features that have more modern replacements:
+
+- use `$state` instead of implicit reactivity (e.g. `let count = 0; count += 1`)
+- use `$derived` and `$effect` instead of `$:` assignments and statements (but only use effects when there is no better solution)
+- use `$props` instead of `export let`, `$$props` and `$$restProps`
+- use `onclick={...}` instead of `on:click={...}`
+- use `{#snippet ...}` and `{@render ...}` instead of `<slot>` and `$$slots` and `<svelte:fragment>`
+- use `<DynamicComponent>` instead of `<svelte:component this={DynamicComponent}>`
+- use `import Self from './ThisComponent.svelte'` and `<Self>` instead of `<svelte:self>`
+- use classes with `$state` fields to share reactivity between components, instead of using stores
+- use `{@attach ...}` instead of `use:action`
+- use clsx-style arrays and objects in `class` attributes, instead of the `class:` directive
+````
+<!-- prettier-ignore-end -->
+
+</details>

documentation/docs/40-skills/10-skills.md

@@ -0,0 +1,11 @@
+---
+title: Overview
+---
+
+This is the list of available skills provided by the Svelte MCP package. Skills are sets of instructions that AI agents can load on-demand to help with specific tasks.
+
+Skills are available in both the Claude Code plugin (installed via the marketplace) and the OpenCode plugin (`@sveltejs/opencode`). They can also be manually installed in your `.claude/skills` or `.opencode/skills` folder.
+
+You can download the latest skills from the [releases page](https://github.com/sveltejs/ai-tools/releases) of the repo, or find them in the [`plugins/svelte/skills`](https://github.com/sveltejs/ai-tools/tree/main/plugins/svelte/skills) folder.
+
+@include .generated/skills.md

documentation/docs/40-skills/index.md

@@ -0,0 +1,3 @@
+---
+title: Skills
+---

documentation/docs/50-subagents/.generated/subagent.md

@@ -0,0 +1,67 @@
+---
+name: svelte-file-editor
+description: Specialized Svelte 5 code editor. MUST BE USED PROACTIVELY when creating, editing, or reviewing any .svelte file or .svelte.ts/.svelte.js module and MUST use the tools from the MCP server or the `svelte-file-editor` skill if they are available. Fetches relevant documentation and validates code using the Svelte MCP server tools.
+---
+
+You are a Svelte 5 expert responsible for writing, editing, and validating Svelte components and modules. You have access to the Svelte MCP server which provides documentation and code analysis tools. Always use the tools from the svelte MCP server to fetch documentation with `get_documentation` and validating the code with `svelte_autofixer`. If the autofixer returns any issue or suggestions try to solve them.
+
+If the MCP tools are not available you can use the `svelte-code-writer` skill to learn how to use the `@sveltejs/mcp` cli to access the same tools.
+
+If the skill is not available you can run `npx @sveltejs/mcp@latest -y --help` to learn how to use it.
+
+## Available MCP Tools
+
+### 1. list-sections
+
+Lists all available Svelte 5 and SvelteKit documentation sections with titles and paths. Use this first to discover what documentation is available.
+
+### 2. get-documentation
+
+Retrieves full documentation for specified sections. Accepts a single section name or an array of section names. Use after `list-sections` to fetch relevant docs for the task at hand.
+
+**Example sections:** `$state`, `$derived`, `$effect`, `$props`, `$bindable`, `snippets`, `routing`, `load functions`
+
+### 3. svelte-autofixer
+
+Analyzes Svelte code and returns suggestions to fix issues. Pass the component code directly to this tool. It will detect common mistakes like:
+
+- Using `$effect` instead of `$derived` for computations
+- Missing cleanup in effects
+- Svelte 4 syntax (`on:click`, `export let`, `<slot>`)
+- Missing keys in `{#each}` blocks
+- And more
+
+## Workflow
+
+When invoked to work on a Svelte file:
+
+### 1. Gather Context (if needed)
+
+If you're uncertain about Svelte 5 syntax or patterns, use the MCP tools:
+
+1. Call `list-sections` to see available documentation
+2. Call `get-documentation` with relevant section names
+
+### 2. Read the Target File
+
+Read the file to understand the current implementation.
+
+### 3. Make Changes
+
+Apply edits following Svelte 5 best practices:
+
+### 4. Validate Changes
+
+After editing, ALWAYS call `svelte-autofixer` with the updated code to check for issues.
+
+### 5. Fix Any Issues
+
+If the autofixer reports problems, fix them and re-validate until no issues remain.
+
+## Output Format
+
+After completing your work, provide:
+
+1. Summary of changes made
+2. Any issues found and fixed by the autofixer
+3. Recommendations for further improvements (if any)

documentation/docs/50-subagents/10-subagent.md

@@ -0,0 +1,20 @@
+---
+title: Overview
+---
+
+Since creating, editing or analyzing a Svelte file is an atomic operation we recommend creating a subagent that your main agent can invoke whenever it needs to interact with a Svelte component. Subagents use a separate context window, allowing them to fetch documentation, iterate with [`svelte-autofixer`](tools#svelte-autofixer) and write to the filesystem without wasting context in the main agent.
+
+Delegation should happen automatically when appropriate, but you can also explicitly request the subagent be used for Svelte-related tasks.
+
+You can write your own or take inspiration from the one available in the [`sveltejs/ai-tools`](https://github.com/sveltejs/ai-tools/tree/main/tools/agents) repository: a specialized subagent called `svelte-file-editor` designed for creating, editing, and reviewing Svelte files.
+
+<details>
+	<summary>View subagent definition</summary>
+
+<!-- prettier-ignore-start -->
+````markdown
+@include .generated/subagent.md
+````
+<!-- prettier-ignore-end -->
+
+</details>

documentation/docs/50-subagents/index.md

@@ -0,0 +1,3 @@
+---
+title: Subagents
+---

documentation/docs/60-plugins/10-claude-plugin.md

@@ -0,0 +1,23 @@
+---
+title: Claude Code
+---
+
+The open source [repository](https://github.com/sveltejs/ai-tools) containing the code for the MCP server is also a Claude Code [plugin marketplace](https://code.claude.com/docs/en/discover-plugins).
+
+The marketplace allows you to install the `svelte` plugin which will give you the remote MCP server, [skills](skills) to instruct the LLM on how to properly write Svelte 5 code, and a specialized agent for editing Svelte files.
+
+If possible, we recommend that you instruct the LLM to execute MCP calls with the agent (you can explicitly mention an agent in your message to delegate work to it) when creating or editing `.svelte` files or `.svelte.ts`/`.svelte.js` modules — this will help save context by handling Svelte-specific tasks more efficiently.
+
+## Installation
+
+To add the repository as a marketplace, launch Claude Code and type the following:
+
+```bash
+/plugin marketplace add sveltejs/ai-tools
+```
+
+Then, install the Svelte plugin:
+
+```bash
+/plugin install svelte
+```

documentation/docs/60-plugins/20-opencode-plugin.md

@@ -0,0 +1,54 @@
+---
+title: OpenCode
+---
+
+OpenCode has a [plugin system](https://opencode.ai/docs/plugins/) that allows developers to add MCP servers, agents and commands programmatically. Svelte has an OpenCode plugin published under `@sveltejs/opencode`.
+
+## Installation
+
+To install the plugin you can edit your [OpenCode config](https://opencode.ai/docs/config/) (either the global or the local one), adding `@sveltejs/opencode` to the list of plugins.
+
+```json
+{
+	"$schema": "https://opencode.ai/config.json",
+	"plugin": ["@sveltejs/opencode"]
+}
+```
+
+That's it! You now have the Svelte [MCP server](mcp), [skills](skills), and the `svelte-file-editor` [subagent](subagent) configured for you.
+
+## Configuration
+
+By default, everything is enabled, but you can configure the plugin by adding a configuration file:
+
+- locally, in `.opencode/svelte.json`
+- globally, in `~/.config/opencode/svelte.json` (or, if you have specified the environment variable, in `$OPENCODE_CONFIG_DIR/svelte.json`)
+
+```json
+{
+	"$schema": "https://svelte.dev/opencode/schema.json",
+	"mcp": {
+		"type": "remote", // or "local" — defaults to remote
+		"enabled": true
+	},
+	"subagent": {
+		"enabled": true,
+		"agents": {
+			"svelte-file-editor": {
+				"model": "<other-model>", // defaults to the same as main agent
+				"temperature": 1, // defaults to unset
+				"top_p": 0.7, // defaults to unset
+				"maxSteps": 20 // defaults to unlimited
+			}
+		}
+	},
+	"skills": {
+		// this can be `true`, or an array of skills to enable
+		// e.g. ["svelte-core-bestpractices"]
+		"enabled": true
+	},
+	"instructions": {
+		"enabled": true
+	}
+}
+```

documentation/docs/60-plugins/30-cursor-plugin.md

@@ -0,0 +1,25 @@
+---
+title: Cursor
+---
+
+Cursor has a [plugin system](https://cursor.com/docs/plugins) that can bundle rules, skills, agents, commands, MCP servers, and hooks.
+
+The Svelte plugin gives you the remote Svelte MCP server, Cursor [skills](skills), an always-on rule that tells the model how to use the Svelte MCP tools correctly, and the `svelte-file-editor` subagent for working on `.svelte` files and `.svelte.ts`/`.svelte.js` modules. The source is available in the [`sveltejs/ai-tools`](https://github.com/sveltejs/ai-tools/tree/main/plugins/cursor/svelte) repo.
+
+## Installation
+
+Install the plugin from the [Cursor Marketplace](https://cursor.com/marketplace/svelte) with the following command:
+
+```
+/add-plugin svelte
+```
+
+Plugins can be installed either for the current project or at user level.
+
+Once installed, Cursor will discover the plugin components automatically:
+
+- the Svelte MCP server is added from the plugin's `.mcp.json`
+- rules and skills appear in Cursor's rules UI
+- the `svelte-file-editor` agent becomes available in chat
+
+> [!NOTE] The Cursor CLI does not support plugins yet. Plugin support in [Cloud Agents](https://cursor.com/docs/cloud-agent) is limited to MCP servers.

documentation/docs/60-plugins/index.md

@@ -0,0 +1,3 @@
+---
+title: Plugins
+---

documentation/docs/index.md

@@ -1,3 +1,3 @@
 ---
-title: MCP
+title: AI
 ---

eslint.config.js

@@ -7,13 +7,20 @@ import { fileURLToPath } from 'node:url';
 import ts from 'typescript-eslint';
 import svelteConfig from './apps/mcp-remote/svelte.config.js';
 import eslint_plugin_import from 'eslint-plugin-import';
+import { configs as pnpm } from 'eslint-plugin-pnpm';
 
 const gitignore_path = fileURLToPath(new URL('./.gitignore', import.meta.url));
 
 export default /** @type {import("eslint").Linter.Config} */ ([
 	includeIgnoreFile(gitignore_path),
 	{
-		ignores: ['.claude/**/*', '.changeset/*'],
+		ignores: [
+			'.claude/**/*',
+			'.changeset/*',
+			'.github/**/*.yml',
+			'.github/**/*.yaml',
+			'**/pnpm-lock.yaml',
+		],
 	},
 	js.configs.recommended,
 	...ts.configs.recommended,
@@ -48,13 +55,17 @@ export default /** @type {import("eslint").Linter.Config} */ ([
 			'import/no-unresolved': 'off', // this doesn't work well with typescript path mapping
 			'import/extensions': [
 				'error',
-				'ignorePackages',
 				{
-					js: 'always',
-					mjs: 'always',
-					cjs: 'always',
-					ts: 'always',
-					svelte: 'always',
+					ignorePackages: true,
+					pattern: {
+						js: 'always',
+						mjs: 'always',
+						cjs: 'always',
+						ts: 'always',
+						svelte: 'always',
+						svg: 'always',
+						json: 'always',
+					},
 				},
 			],
 		},
@@ -70,4 +81,16 @@ export default /** @type {import("eslint").Linter.Config} */ ([
 			},
 		},
 	},
+	{
+		name: 'pnpm/exclude-some-rules',
+		files: ['**/*.json', '**/*.yaml', '**/*.yml', 'pnpm-workspace.yaml'],
+		rules: {
+			'@typescript-eslint/naming-convention': 'off',
+			'@typescript-eslint/no-unused-vars': 'off',
+			'@typescript-eslint/no-unused-expressions': 'off',
+			'func-style': 'off',
+		},
+	},
+	...pnpm.json,
+	...pnpm.yaml,
 ]);

package.json

@@ -3,7 +3,7 @@
 	"version": "0.0.1",
 	"description": "The official Svelte MCP server implementation",
 	"type": "module",
-	"packageManager": "pnpm@10.19.0",
+	"packageManager": "pnpm@10.28.2",
 	"scripts": {
 		"build": "pnpm -r run build",
 		"dev": "pnpm --filter @sveltejs/mcp-remote run dev",
@@ -12,15 +12,28 @@
 		"format": "prettier --write .",
 		"lint": "prettier --check . && eslint .",
 		"lint:fix": "prettier --write . && eslint . --fix",
+		"lint:inspect": "pnpm dlx @eslint/config-inspector",
+		"node:inspect": "pnpm dlx node-modules-inspector",
 		"test:unit": "vitest",
 		"test": "npm run test:unit -- --run",
 		"test:watch": "npm run test:unit -- --watch",
 		"inspect": "pnpm mcp-inspector",
+		"generate-opencode-jsonschema": "pnpm --filter @sveltejs/opencode run generate-schema",
 		"generate-summaries": "pnpm --filter @sveltejs/mcp-server run generate-summaries",
 		"generate-prompt-docs": "node --import node-resolve-ts/register scripts/update-docs-prompts.ts",
+		"generate-skill-docs": "node --import node-resolve-ts/register scripts/update-docs-skills.ts",
+		"generate-subagent-docs": "cp ./tools/agents/svelte-file-editor.md ./documentation/docs/50-subagents/.generated/subagent.md",
 		"debug:generate-summaries": "pnpm --filter @sveltejs/mcp-server run debug:generate-summaries",
 		"release": "pnpm --filter @sveltejs/mcp run build && changeset publish",
-		"changeset:version": "changeset version && pnpm --filter @sveltejs/mcp run update:version && git add --all"
+		"changeset:version": "changeset version && pnpm --filter @sveltejs/mcp run update:version && git add --all",
+		"sync-plugins": "pnpm sync-claude-plugin && pnpm sync-cursor-plugin && pnpm sync-opencode-plugin && pnpm bump-plugin-versions",
+		"sync-claude-plugin": "node scripts/sync-claude-plugin.ts",
+		"sync-cursor-plugin": "node scripts/sync-cursor-plugin.ts",
+		"sync-opencode-plugin": "node scripts/sync-opencode-plugin.ts && pnpm generate-opencode-jsonschema",
+		"bump-plugin-versions": "node scripts/bump-plugin-versions.ts",
+		"postbump-plugin-versions": "pnpm format",
+		"resolve-references": "node scripts/resolve-references.ts",
+		"postresolve-references": "pnpm format"
 	},
 	"keywords": [
 		"svelte",
@@ -30,27 +43,24 @@
 	],
 	"private": true,
 	"devDependencies": {
-		"@changesets/cli": "^2.29.7",
-		"@eslint/compat": "^1.3.2",
-		"@eslint/js": "^9.36.0",
-		"@modelcontextprotocol/inspector": "^0.17.0",
-		"@svitejs/changesets-changelog-github-compact": "^1.2.0",
-		"eslint": "^9.36.0",
-		"eslint-config-prettier": "^10.0.1",
-		"eslint-plugin-import": "^2.32.0",
-		"eslint-plugin-svelte": "^3.12.3",
-		"globals": "^16.0.0",
-		"node-resolve-ts": "^1.0.2",
-		"prettier": "^3.4.2",
-		"prettier-plugin-svelte": "^3.3.3",
-		"publint": "^0.3.13",
-		"typescript": "^5.0.0",
-		"typescript-eslint": "^8.44.1",
-		"vitest": "^4.0.0"
-	},
-	"pnpm": {
-		"onlyBuiltDependencies": [
-			"esbuild"
-		]
+		"@changesets/cli": "catalog:tooling",
+		"@eslint/compat": "catalog:lint",
+		"@eslint/js": "catalog:lint",
+		"@modelcontextprotocol/inspector": "catalog:ai",
+		"@sveltejs/adapter-vercel": "catalog:svelte",
+		"@svitejs/changesets-changelog-github-compact": "catalog:tooling",
+		"eslint": "catalog:lint",
+		"eslint-config-prettier": "catalog:lint",
+		"eslint-plugin-import": "catalog:lint",
+		"eslint-plugin-pnpm": "catalog:lint",
+		"eslint-plugin-svelte": "catalog:lint",
+		"globals": "catalog:lint",
+		"node-resolve-ts": "catalog:tooling",
+		"prettier": "catalog:lint",
+		"prettier-plugin-svelte": "catalog:lint",
+		"publint": "catalog:tooling",
+		"typescript": "catalog:tooling",
+		"typescript-eslint": "catalog:lint",
+		"vitest": "catalog:tooling"
 	}
 }

packages/mcp-schema/package.json

@@ -1,19 +0,0 @@
-{
-	"name": "@sveltejs/mcp-schema",
-	"version": "0.0.1",
-	"private": true,
-	"description": "",
-	"main": "index.js",
-	"exports": {
-		".": "./src/index.js",
-		"./utils": "./src/utils.js",
-		"./schema": "./src/schema.js"
-	},
-	"keywords": [],
-	"author": "",
-	"license": "ISC",
-	"type": "module",
-	"dependencies": {
-		"drizzle-orm": "^0.44.0"
-	}
-}

packages/mcp-schema/src/index.js

@@ -1,8 +0,0 @@
-/**
- * @import * as schema from './schema.js'
- */
-export * from './schema.js';
-
-/**
- * @typedef {typeof schema} Schema
- */

packages/mcp-schema/src/schema.js

@@ -1,82 +0,0 @@
-import { integer, sqliteTable, text } from 'drizzle-orm/sqlite-core';
-import { float_32_array } from './utils.js';
-
-/**
- * NOTE: if you modify a schema adding a vector column you need to manually add this
- *
- * CREATE INDEX IF NOT EXISTS name_of_the_index
- * ON `name_of_the_table` (
- * 	libsql_vector_idx(name_of_the_column, 'metric=cosine')
- * )
- *
- * to the generated migration file
- */
-
-export const distillations = sqliteTable('distillations', {
-	id: integer('id').primaryKey(),
-	preset_name: text('preset_name').notNull(),
-	version: text('version').notNull(),
-	content: text('content').notNull(),
-	size_kb: integer('size_kb').notNull(),
-	document_count: integer('document_count').notNull(),
-	distillation_job_id: integer('distillation_job_id').references(() => distillation_jobs.id),
-	created_at: integer('created_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-});
-
-export const distillation_jobs = sqliteTable('distillation_jobs', {
-	id: integer('id').primaryKey(),
-	preset_name: text('preset_name').notNull(),
-	batch_id: text('batch_id'),
-	status: text('status', { enum: ['pending', 'processing', 'completed', 'failed'] }).notNull(),
-	model_used: text('model_used').notNull(),
-	total_files: integer('total_files').notNull(),
-	processed_files: integer('processed_files').notNull().default(0),
-	successful_files: integer('successful_files').notNull().default(0),
-	minimize_applied: integer('minimize_applied', { mode: 'boolean' }).notNull().default(false),
-	total_input_tokens: integer('total_input_tokens').notNull().default(0),
-	total_output_tokens: integer('total_output_tokens').notNull().default(0),
-	started_at: integer('started_at', { mode: 'timestamp' }),
-	completed_at: integer('completed_at', { mode: 'timestamp' }),
-	error_message: text('error_message'),
-	metadata: text('metadata', { mode: 'json' }).notNull().default({}),
-	created_at: integer('created_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-	updated_at: integer('updated_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-});
-
-export const content = sqliteTable('content', {
-	id: integer('id').primaryKey(),
-	path: text('path').notNull(),
-	filename: text('filename').notNull(),
-	content: text('content').notNull(),
-	size_bytes: integer('size_bytes').notNull(),
-	embeddings: float_32_array('embeddings', { dimensions: 1024 }),
-	metadata: text('metadata', { mode: 'json' }).notNull().default({}),
-	created_at: integer('created_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-	updated_at: integer('updated_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-});
-
-export const content_distilled = sqliteTable('content_distilled', {
-	id: integer('id').primaryKey(),
-	path: text('path').notNull(),
-	filename: text('filename').notNull(),
-	content: text('content').notNull(),
-	size_bytes: integer('size_bytes').notNull(),
-	embeddings: float_32_array('embeddings', { dimensions: 1024 }),
-	metadata: text('metadata', { mode: 'json' }).notNull().default({}),
-	created_at: integer('created_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-	updated_at: integer('updated_at', { mode: 'timestamp' })
-		.notNull()
-		.$defaultFn(() => new Date()),
-});

packages/mcp-schema/src/utils.js

@@ -1,65 +0,0 @@
-/**
- * @import { Column } from 'drizzle-orm';
- */
-import { sql } from 'drizzle-orm';
-import { customType } from 'drizzle-orm/sqlite-core';
-
-/**
- * Helper function to convert an array of embeddings into a format that can be inserted into a LibSQL vector column.
- * @param {number[]} arr The embeddings array.
- */
-export function vector(arr) {
-	return sql`vector32(${JSON.stringify(arr)})`;
-}
-
-/**
- * Helper function to calculate the distance between a vector column and an array of embeddings and return it as a columns.
- * @param {Column} column The drizzle column representing the vector.
- * @param {number} arr The embeddings array.
- * @param {string} as The name of the returned column. Default is 'distance'.
- *
- * @example
- * await db.select({
- *   id: vector_table.id,
- *   text: vector_table.text,
- *   distance: distance(vector_table.vector, await get_embeddings(sentence)),
- * })
- * .from(vector_table)
- * .orderBy(sql`distance`)
- * .execute();
- */
-export function distance(column, arr, as = 'distance') {
-	return /** @type {typeof sql<number>} */ (
-		sql
-	)`CASE ${column} ISNULL WHEN 1 THEN 1 ELSE vector_distance_cos(${column}, vector32(${JSON.stringify(arr)})) END`.as(
-		as,
-	);
-}
-
-/**
- * Custom drizzle type to use the LibSQL vector column type.
- */
-export const float_32_array = /** @type {typeof customType<{
-	data: number[];
-	config: { dimensions: number };
-	configRequired: true;
-	driverData: Buffer;
-}>} */ (customType)({
-	dataType(config) {
-		return `F32_BLOB(${config.dimensions})`;
-	},
-	/**
-	 * @param {Buffer} value
-	 */
-	fromDriver(value) {
-		return Array.from(new Float32Array(value.buffer));
-	},
-	/**
-	 *
-	 * @param {number[]} value
-	 * @returns
-	 */
-	toDriver(value) {
-		return vector(value);
-	},
-});

packages/mcp-server/package.json

@@ -14,33 +14,31 @@
 		"debug:generate-summaries": "DEBUG_MODE=1 node scripts/generate-summaries.ts --experimental-strip-types"
 	},
 	"exports": {
-		".": "./src/index.ts"
-	},
-	"peerDependencies": {
-		"drizzle-orm": "^0.44.0"
+		".": "./src/index.ts",
+		"./handlers": "./src/mcp/handlers/tools/handlers.ts"
 	},
 	"dependencies": {
-		"@mcp-ui/server": "^5.12.0",
-		"@sveltejs/mcp-schema": "workspace:^",
-		"@tmcp/adapter-valibot": "^0.1.4",
-		"@typescript-eslint/parser": "^8.44.0",
-		"eslint": "^9.36.0",
-		"eslint-plugin-svelte": "^3.12.3",
-		"svelte": "^5.39.2",
-		"svelte-eslint-parser": "^1.3.2",
-		"tmcp": "^1.15.3",
-		"ts-blank-space": "^0.6.2",
-		"typescript-eslint": "^8.44.0",
-		"valibot": "^1.1.0",
-		"vitest": "^4.0.0",
-		"zimmerframe": "^1.1.4"
+		"@mcp-ui/server": "catalog:ai",
+		"@tmcp/adapter-valibot": "catalog:tmcp",
+		"@tmcp/transport-in-memory": "catalog:tmcp",
+		"@typescript-eslint/parser": "catalog:lint",
+		"eslint": "catalog:lint",
+		"eslint-plugin-svelte": "catalog:lint",
+		"svelte": "catalog:svelte",
+		"svelte-eslint-parser": "catalog:lint",
+		"tmcp": "catalog:tmcp",
+		"ts-blank-space": "catalog:tooling",
+		"typescript-eslint": "catalog:lint",
+		"valibot": "catalog:tooling",
+		"vitest": "catalog:tooling",
+		"zimmerframe": "catalog:tooling"
 	},
 	"devDependencies": {
-		"@anthropic-ai/sdk": "^0.67.0",
-		"@sveltejs/kit": "^2.42.2",
-		"@types/eslint-scope": "^8.3.2",
-		"@types/estree": "^1.0.8",
-		"@typescript-eslint/types": "^8.44.0",
-		"dotenv": "^17.2.3"
+		"@anthropic-ai/sdk": "catalog:ai",
+		"@sveltejs/kit": "catalog:svelte",
+		"@types/eslint-scope": "catalog:lint",
+		"@types/estree": "catalog:tooling",
+		"@typescript-eslint/types": "catalog:lint",
+		"dotenv": "catalog:tooling"
 	}
 }

packages/mcp-server/src/mcp/autofixers/add-autofixers-issues.test.ts

@@ -344,87 +344,111 @@ describe('add_autofixers_issues', () => {
 	});
 
 	describe('imported_runes', () => {
-		describe.each([{ source: 'svelte' }, { source: 'svelte/runes' }])(
-			'from "$source"',
-			({ source }) => {
-				describe.each(dollarless_runes)('single import ($rune)', ({ rune }) => {
-					it(`should add suggestions when importing '${rune}' from '${source}'`, () => {
-						const content = run_autofixers_on_code(`
+		describe.each([
+			{ source: 'svelte' },
+			{ source: 'svelte/runes' },
+			{ source: '@sveltejs/runes' },
+			{ source: '@sveltejs/vite-plugin-svelte' },
+		])('from "$source"', ({ source }) => {
+			describe.each(dollarless_runes)('single import ($rune)', ({ rune }) => {
+				it(`should add suggestions when importing '${rune}' from '${source}'`, () => {
+					const content = run_autofixers_on_code(`
 					<script>
 						import { ${rune} } from '${source}';
 					</script>`);
 
-						expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
-						expect(content.suggestions).toContain(
-							`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
-						);
-					});
+					expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
+					expect(content.suggestions).toContain(
+						`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
+					);
+				});
 
-					it(`should add suggestions when importing "${rune}" as the default export from '${source}'`, () => {
-						const content = run_autofixers_on_code(`
+				it(`should add suggestions when importing "${rune}" as the default export from '${source}'`, () => {
+					const content = run_autofixers_on_code(`
 					<script>
 						import ${rune} from '${source}';
 					</script>`);
 
-						expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
-						expect(content.suggestions).toContain(
-							`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
-						);
-					});
+					expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
+					expect(content.suggestions).toContain(
+						`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
+					);
+				});
 
-					it(`should add suggestions when importing '${rune}' as the namespace export from '${source}'`, () => {
-						const content = run_autofixers_on_code(`
+				it(`should add suggestions when importing '${rune}' as the namespace export from '${source}'`, () => {
+					const content = run_autofixers_on_code(`
 					<script>
 						import * as ${rune} from '${source}';
 					</script>`);
 
-						expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
-						expect(content.suggestions).toContain(
-							`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
-						);
-					});
+					expect(content.suggestions.length).toBeGreaterThanOrEqual(1);
+					expect(content.suggestions).toContain(
+						`You are importing "${rune}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
+					);
 				});
+			});
 
-				it(`should add suggestions when importing multiple runes from '${source}'`, () => {
-					const content = run_autofixers_on_code(`
+			it(`should add suggestions when importing multiple runes from '${source}'`, () => {
+				const content = run_autofixers_on_code(`
 					<script>
 						import { onMount, state, effect } from '${source}';
 					</script>`);
 
-					expect(content.suggestions.length).toBeGreaterThanOrEqual(2);
-					expect(content.suggestions).toContain(
-						`You are importing "state" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$state" directly.`,
-					);
-					expect(content.suggestions).toContain(
-						`You are importing "effect" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$effect" directly.`,
-					);
-				});
+				expect(content.suggestions.length).toBeGreaterThanOrEqual(2);
+				expect(content.suggestions).toContain(
+					`You are importing "state" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$state" directly.`,
+				);
+				expect(content.suggestions).toContain(
+					`You are importing "effect" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$effect" directly.`,
+				);
+			});
 
-				it(`should not add suggestions when importing other identifiers from '${source}'`, () => {
-					const content = run_autofixers_on_code(`
+			it(`should not add suggestions when importing other identifiers from '${source}'`, () => {
+				const content = run_autofixers_on_code(`
 					<script>
 						import { onMount } from '${source}';
 					</script>`);
 
-					expect(content.suggestions).not.toContain(
-						`You are importing "onMount" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$onMount" directly.`,
-					);
-				});
-			},
-		);
+				expect(content.suggestions).not.toContain(
+					`You are importing "onMount" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$onMount" directly.`,
+				);
+			});
+		});
 
 		describe.each(dollarless_runes)('importing $rune from external lib', ({ rune }) => {
-			it(`should not add suggestions when importing from packages that are not svelte`, () => {
+			it(`should not add suggestions when importing from packages whose name doesn't contain svelte`, () => {
+				const content = run_autofixers_on_code(`
+				<script>
+					import { ${rune} } from 'something-something';
+				</script>`);
+
+				expect(content.suggestions).not.toContain(
+					`You are importing "${rune}" from "something-something". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
+				);
+			});
+
+			it(`should add suggestions with a different hint when importing from packages whose name contains svelte but it's not official`, () => {
 				const content = run_autofixers_on_code(`
 				<script>
 					import { ${rune} } from 'svelte-something-something';
 				</script>`);
 
-				expect(content.suggestions).not.toContain(
-					`You are importing "${rune}" from "svelte-something-something". This is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly.`,
+				expect(content.suggestions).toContain(
+					`You are importing "${rune}" from "svelte-something-something". If you are trying to import runes to use them this is not necessary, all runes are globally available. Please remove this import and use "$${rune}" directly. If you are importing the function from a separate library ignore this suggestion.`,
 				);
 			});
 		});
+
+		it('should not add the imported_runes suggestion when importing derived from svelte/store', () => {
+			const content = run_autofixers_on_code(`
+			<script>
+				import { derived } from 'svelte/store';
+			</script>`);
+
+			expect(content.suggestions).not.toContain(
+				'You are importing "derived" from "svelte/store". This is not necessary, all runes are globally available. Please remove this import and use "$derived" directly.',
+			);
+		});
 	});
 
 	describe('derived_with_function', () => {
@@ -751,5 +775,16 @@ describe('add_autofixers_issues', () => {
 				`You are reading the stateful variable "$x" with a "$" prefix. Stateful variables are not stores and should be read without the "$". Please read it as a normal variable "x"`,
 			);
 		});
+
+		// https://github.com/sveltejs/ai-tools/issues/200
+		it('should not crash when reading $-prefixed identifier for variable initialized with member expression', () => {
+			expect(() =>
+				run_autofixers_on_code(`<div>{$x}</div>
+
+					<script>
+					const x = foo.bar;
+					</script>`),
+			).not.toThrow();
+		});
 	});
 });

packages/mcp-server/src/mcp/autofixers/add-autofixers-issues.ts

@@ -8,6 +8,7 @@ export function add_autofixers_issues(
 	code: string,
 	desired_svelte_version: number,
 	filename = 'Component.svelte',
+	async = false,
 ) {
 	const parsed = parse(code, filename);
 
@@ -15,7 +16,7 @@ export function add_autofixers_issues(
 	for (const autofixer of Object.values(autofixers)) {
 		walk(
 			parsed.ast as unknown as Node,
-			{ output: content, parsed, desired_svelte_version },
+			{ output: content, parsed, desired_svelte_version, async },
 			autofixer,
 		);
 	}

packages/mcp-server/src/mcp/autofixers/add-compile-issues.ts

@@ -7,6 +7,7 @@ export function add_compile_issues(
 	code: string,
 	desired_svelte_version: number,
 	filename = 'Component.svelte',
+	async = false,
 ) {
 	let compile = compile_component;
 	const extension = extname(filename);
@@ -27,6 +28,7 @@ export function add_compile_issues(
 		filename: filename || 'Component.svelte',
 		generate: false,
 		runes: desired_svelte_version >= 5,
+		experimental: { async },
 	});
 
 	for (const warning of compilation_result.warnings) {

packages/mcp-server/src/mcp/autofixers/add-eslint-issues.ts

@@ -35,6 +35,7 @@ function base_config(svelte_config: Config): ESLint.Options['baseConfig'] {
 				'svelte/prefer-writable-derived': 'warn',
 				'svelte/require-event-dispatcher-types': 'warn',
 				'svelte/require-store-reactive-access': 'warn',
+				'svelte/no-inspect': 'off',
 			},
 
 			languageOptions: {
@@ -51,7 +52,7 @@ function base_config(svelte_config: Config): ESLint.Options['baseConfig'] {
 	];
 }
 
-function get_linter(version: number) {
+function get_linter(version: number, async = false) {
 	if (version < 5) {
 		return (svelte_4_linter ??= new ESLint({
 			overrideConfigFile: true,
@@ -67,6 +68,7 @@ function get_linter(version: number) {
 		baseConfig: base_config({
 			compilerOptions: {
 				runes: true,
+				experimental: { async },
 			},
 		}),
 	}));
@@ -77,8 +79,9 @@ export async function add_eslint_issues(
 	code: string,
 	desired_svelte_version: number,
 	filename = 'Component.svelte',
+	async = false,
 ) {
-	const eslint = get_linter(desired_svelte_version);
+	const eslint = get_linter(desired_svelte_version, async);
 	const results = await eslint.lintText(code, { filePath: filename || './Component.svelte' });
 
 	for (const message of results[0]?.messages ?? []) {

packages/mcp-server/src/mcp/autofixers/visitors/imported-runes.ts

@@ -3,10 +3,20 @@ import type { Autofixer } from './index.js';
 
 const dollarless_runes = base_runes.map((r) => r.replace('$', ''));
 
+function should_suggest_for_source(source: string, rune: string) {
+	if (!source.includes('svelte')) {
+		return false;
+	}
+	if (source === 'svelte/store' && rune === 'derived') {
+		return false;
+	}
+	return true;
+}
+
 export const imported_runes: Autofixer = {
 	ImportDeclaration(node, { state, next }) {
 		const source = (node.source.value || node.source.raw?.slice(1, -1))?.toString();
-		if (source && (source === 'svelte' || source.startsWith('svelte/'))) {
+		if (source) {
 			for (const specifier of node.specifiers) {
 				const id =
 					specifier.type === 'ImportDefaultSpecifier'
@@ -16,10 +26,25 @@ export const imported_runes: Autofixer = {
 							: specifier.type === 'ImportSpecifier'
 								? specifier.imported
 								: null;
-				if (id && id.type === 'Identifier' && dollarless_runes.includes(id.name)) {
-					state.output.suggestions.push(
-						`You are importing "${id.name}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${id.name}" directly.`,
-					);
+				if (
+					id &&
+					id.type === 'Identifier' &&
+					dollarless_runes.includes(id.name) &&
+					should_suggest_for_source(source, id.name)
+				) {
+					if (
+						source === 'svelte' ||
+						source.startsWith('svelte/') ||
+						source.startsWith('@sveltejs')
+					) {
+						state.output.suggestions.push(
+							`You are importing "${id.name}" from "${source}". This is not necessary, all runes are globally available. Please remove this import and use "$${id.name}" directly.`,
+						);
+					} else {
+						state.output.suggestions.push(
+							`You are importing "${id.name}" from "${source}". If you are trying to import runes to use them this is not necessary, all runes are globally available. Please remove this import and use "$${id.name}" directly. If you are importing the function from a separate library ignore this suggestion.`,
+						);
+					}
 				}
 			}
 		}

packages/mcp-server/src/mcp/autofixers/visitors/index.ts

@@ -7,6 +7,7 @@ export type AutofixerState = {
 	output: { issues: string[]; suggestions: string[] };
 	parsed: ParseResult;
 	desired_svelte_version: number;
+	async?: boolean;
 };
 
 export type Autofixer = Visitors<Node | AST.SvelteNode, AutofixerState>;

packages/mcp-server/src/mcp/handlers/prompts/svelte-task.ts

@@ -2,19 +2,22 @@ import type { SvelteMcp } from '../../index.js';
 import * as v from 'valibot';
 import { format_sections_list } from '../../utils.js';
 import { icons } from '../../icons/index.js';
+import { prompt } from 'tmcp/utils';
 
 /**
  *  Function that actually generates the prompt string. You can use this in the MCP server handler to generate the prompt, it can accept arguments
  *  if needed (it will always be invoked manually so it's up to you to provide the arguments).
  */
 function svelte_task(available_docs: string, task: string) {
-	return `You are a Svelte expert tasked to build components and utilities for Svelte developers. If you need documentation for anything related to Svelte you can invoke the tool \`get_documentation\` with one of the following paths:
+	return `You are a Svelte expert tasked to build components and utilities for Svelte developers. If you need documentation for anything related to Svelte you can invoke the tool \`get-documentation\` with one of the following paths. However: before invoking the \`get-documentation\` tool, try to answer the users query using your own knowledge and the \`svelte-autofixer\` tool. Be mindful of how many section you request, since it is token-intensive!
 <available-docs>
 
 ${available_docs}
 
 </available-docs>
 
+These are the available documentation sections that \`list-sections\` will return, you do not need to call it again.
+
 Every time you write a Svelte component or a Svelte module you MUST invoke the \`svelte-autofixer\` tool providing the code. The tool will return a list of issues or suggestions. If there are any issues or suggestions you MUST fix them and call the tool again with the updated code. You MUST keep doing this until the tool returns no issues or suggestions. Only then you can return the code to the user.
 
 This is the task you will work on:
@@ -68,19 +71,12 @@ export function setup_svelte_task(server: SvelteMcp) {
 			icons,
 		},
 		async ({ task }) => {
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(server.ctx.sessionId, 'svelte-task');
+			}
 			const available_docs = await format_sections_list();
 
-			return {
-				messages: [
-					{
-						role: 'user',
-						content: {
-							type: 'text',
-							text: svelte_task(available_docs, task),
-						},
-					},
-				],
-			};
+			return prompt.text(svelte_task(available_docs, task));
 		},
 	);
 }

packages/mcp-server/src/mcp/handlers/resources/doc-section.ts

@@ -1,6 +1,7 @@
 import type { SvelteMcp } from '../../index.js';
 import { get_sections, fetch_with_timeout } from '../../utils.js';
 import { icons } from '../../icons/index.js';
+import { resource } from 'tmcp/utils';
 
 export async function list_sections(server: SvelteMcp) {
 	const sections = await get_sections();
@@ -46,21 +47,20 @@ export async function list_sections(server: SvelteMcp) {
 			icons,
 		},
 		async (uri, { slug }) => {
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(
+					server.ctx.sessionId,
+					'svelte-doc-section',
+					Array.isArray(slug) ? slug.join(',') : slug,
+				);
+			}
 			const section = sections.find((section) => {
 				return slug === section.slug;
 			});
 			if (!section) throw new Error(`Section not found: ${slug}`);
 			const response = await fetch_with_timeout(section.url);
 			const content = await response.text();
-			return {
-				contents: [
-					{
-						uri,
-						type: 'text',
-						text: content,
-					},
-				],
-			};
+			return resource.text(uri, content);
 		},
 	);
 }

packages/mcp-server/src/mcp/handlers/tools/get-documentation.ts

@@ -3,115 +3,179 @@ import * as v from 'valibot';
 import { get_sections, fetch_with_timeout, format_sections_list } from '../../utils.js';
 import { SECTIONS_LIST_INTRO, SECTIONS_LIST_OUTRO } from './prompts.js';
 import { icons } from '../../icons/index.js';
+import { tool } from 'tmcp/utils';
+
+const get_documentation_schema = v.object({
+	section: v.pipe(
+		v.union([v.string(), v.array(v.string())]),
+		v.description(
+			'The section name(s) to retrieve. Can search by title (e.g., "$state", "load functions") or file path (e.g., "cli/overview"). Supports single string and array of strings',
+		),
+	),
+});
+
+export async function get_documentation_handler({
+	section,
+}: v.InferInput<typeof get_documentation_schema>) {
+	let sections: string[];
+
+	if (Array.isArray(section)) {
+		sections = section.filter((s): s is string => typeof s === 'string');
+	} else if (
+		typeof section === 'string' &&
+		section.trim().startsWith('[') &&
+		section.trim().endsWith(']')
+	) {
+		try {
+			const parsed = JSON.parse(section);
+			if (Array.isArray(parsed)) {
+				sections = parsed.filter((s): s is string => typeof s === 'string');
+			} else {
+				sections = [section];
+			}
+		} catch {
+			sections = [section];
+		}
+	} else if (typeof section === 'string') {
+		sections = [section];
+	} else {
+		sections = [];
+	}
+
+	const available_sections = await get_sections();
+
+	const settled_results = await Promise.allSettled(
+		sections.map(async (requested_section) => {
+			const matched_section = available_sections.find(
+				(s) =>
+					s.title.toLowerCase() === requested_section.toLowerCase() ||
+					s.slug === requested_section ||
+					s.url === requested_section,
+			);
+
+			if (matched_section) {
+				try {
+					const response = await fetch_with_timeout(matched_section.url);
+					if (response.ok) {
+						const content = await response.text();
+						return { success: true, content: `## ${matched_section.title}\n\n${content}` };
+					} else {
+						return {
+							success: false,
+							content: `## ${matched_section.title}\n\nError: Could not fetch documentation (HTTP ${response.status})`,
+						};
+					}
+				} catch (error) {
+					return {
+						success: false,
+						content: `## ${matched_section.title}\n\nError: Failed to fetch documentation - ${error}`,
+					};
+				}
+			} else {
+				return {
+					success: false,
+					content: `## ${requested_section}\n\nError: Section not found.`,
+				};
+			}
+		}),
+	);
+
+	const results = settled_results.map((result) => {
+		if (result.status === 'fulfilled') {
+			return result.value;
+		} else {
+			return {
+				success: false,
+				content: `Error: Couldn't fetch - ${result.reason}`,
+			};
+		}
+	});
+
+	const successes = results.filter((r) => r.success);
+	const failed_sections = sections.filter(
+		(s) =>
+			!available_sections.some(
+				(a) => a.title.toLowerCase() === s.toLowerCase() || a.slug === s || a.url === s,
+			),
+	);
+
+	if (successes.length > 0 && failed_sections.length === 0) {
+		return successes.map((r) => r.content).join('\n\n---\n\n');
+	}
+
+	const parts: string[] = [];
+
+	if (successes.length > 0) {
+		parts.push(successes.map((r) => r.content).join('\n\n---\n\n'));
+	}
+
+	const fuzzy_results = failed_sections.map((requested) => {
+		const lower = requested.toLowerCase();
+		const matches = available_sections.filter(
+			(a) =>
+				a.title.toLowerCase().includes(lower) ||
+				a.slug.includes(lower) ||
+				lower.includes(a.slug.split('/').pop() ?? '') ||
+				a.use_cases.toLowerCase().includes(lower),
+		);
+		return { requested, matches };
+	});
+
+	const has_fuzzy = fuzzy_results.some((r) => r.matches.length > 0);
+
+	// Full list only when no successes and no fuzzy matches
+	if (successes.length === 0 && !has_fuzzy) {
+		const formatted_sections = await format_sections_list();
+		parts.push(`${SECTIONS_LIST_INTRO}\n\n${formatted_sections}\n\n${SECTIONS_LIST_OUTRO}`);
+	}
+
+	// Similar results then errors
+	for (const { requested, matches } of fuzzy_results) {
+		if (matches.length > 0) {
+			const match_list = matches.map((m) => `- title: ${m.title}, section: ${m.slug}`).join('\n');
+			parts.push(
+				`${matches.length} similar result${matches.length > 1 ? 's' : ''} for "${requested}":\n${match_list}`,
+			);
+		}
+		parts.push(`Section not found: "${requested}".`);
+	}
+
+	return parts.join('\n\n---\n\n');
+}
 
 export function get_documentation(server: SvelteMcp) {
 	server.tool(
 		{
 			name: 'get-documentation',
 			description:
-				'Retrieves full documentation content for Svelte 5 or SvelteKit sections. Supports flexible search by title (e.g., "$state", "routing") or file path (e.g., "cli/overview"). Can accept a single section name or an array of sections. Before running this, make sure to analyze the users query, as well as the output from list-sections (which should be called first). Then ask for ALL relevant sections the user might require. For example, if the user asks to build anything interactive, you will need to fetch all relevant runes, and so on.',
-			schema: v.object({
-				section: v.pipe(
-					v.union([v.string(), v.array(v.string())]),
-					v.description(
-						'The section name(s) to retrieve. Can search by title (e.g., "$state", "load functions") or file path (e.g., "cli/overview"). Supports single string and array of strings',
-					),
-				),
-			}),
+				'Retrieves full documentation content for Svelte 5 or SvelteKit sections. Supports flexible search by title (e.g., "$state", "routing") or file path (e.g., "cli/overview"). Can accept a single section name or an array of sections. Before running this, make sure to analyze the users query, as well as the output from list-sections (which should be called first). Then ask for ALL relevant sections the user might require. For example, if the user asks to build anything interactive, you will need to fetch all relevant runes, and so on. Before calling this tool, try to implement Svelte components using your own knowledge and the `svelte-autofixer` tool, since calling this tool is token intensive.',
+			schema: get_documentation_schema,
+			annotations: {
+				title: 'Get Documentation',
+				destructiveHint: false,
+				readOnlyHint: true,
+				openWorldHint: false,
+			},
 			icons,
 		},
 		async ({ section }) => {
-			let sections: string[];
-
-			if (Array.isArray(section)) {
-				sections = section.filter((s): s is string => typeof s === 'string');
-			} else if (
-				typeof section === 'string' &&
-				section.trim().startsWith('[') &&
-				section.trim().endsWith(']')
-			) {
-				try {
-					const parsed = JSON.parse(section);
-					if (Array.isArray(parsed)) {
-						sections = parsed.filter((s): s is string => typeof s === 'string');
-					} else {
-						sections = [section];
-					}
-				} catch {
-					sections = [section];
-				}
-			} else if (typeof section === 'string') {
-				sections = [section];
-			} else {
-				sections = [];
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(server.ctx.sessionId, 'get-documentation');
 			}
-
-			const available_sections = await get_sections();
-
-			const settled_results = await Promise.allSettled(
-				sections.map(async (requested_section) => {
-					const matched_section = available_sections.find(
-						(s) =>
-							s.title.toLowerCase() === requested_section.toLowerCase() ||
-							s.slug === requested_section ||
-							s.url === requested_section,
+			try {
+				const content = await get_documentation_handler({ section });
+				return tool.text(content);
+			} catch (e) {
+				const error = e as Error;
+				if (server.ctx.sessionId && server.ctx.custom?.track) {
+					await server.ctx.custom?.track?.(
+						server.ctx.sessionId,
+						'get-documentation-error',
+						error.message,
 					);
-
-					if (matched_section) {
-						try {
-							const response = await fetch_with_timeout(matched_section.url);
-							if (response.ok) {
-								const content = await response.text();
-								return { success: true, content: `## ${matched_section.title}\n\n${content}` };
-							} else {
-								return {
-									success: false,
-									content: `## ${matched_section.title}\n\nError: Could not fetch documentation (HTTP ${response.status})`,
-								};
-							}
-						} catch (error) {
-							return {
-								success: false,
-								content: `## ${matched_section.title}\n\nError: Failed to fetch documentation - ${error}`,
-							};
-						}
-					} else {
-						return {
-							success: false,
-							content: `## ${requested_section}\n\nError: Section not found.`,
-						};
-					}
-				}),
-			);
-
-			const results = settled_results.map((result) => {
-				if (result.status === 'fulfilled') {
-					return result.value;
-				} else {
-					return {
-						success: false,
-						content: `Error: Couldn't fetch - ${result.reason}`,
-					};
 				}
-			});
-
-			const has_any_success = results.some((result) => result.success);
-			let final_text = results.map((r) => r.content).join('\n\n---\n\n');
-
-			if (!has_any_success) {
-				const formatted_sections = await format_sections_list();
-
-				final_text += `\n\n---\n\n${SECTIONS_LIST_INTRO}\n\n${formatted_sections}\n\n${SECTIONS_LIST_OUTRO}`;
+				return tool.error(error.message);
 			}
-
-			return {
-				content: [
-					{
-						type: 'text',
-						text: final_text,
-					},
-				],
-			};
 		},
 	);
 }

packages/mcp-server/src/mcp/handlers/tools/handlers.ts

@@ -0,0 +1,4 @@
+export { get_documentation_handler } from './get-documentation.js';
+export { list_sections_handler } from './list-sections.js';
+export { svelte_autofixer_handler } from './svelte-autofixer.js';
+export { playground_link_handler } from './playground-link.js';

packages/mcp-server/src/mcp/handlers/tools/index.ts

@@ -1,4 +1,4 @@
-export * from './get-documentation.js';
-export * from './list-sections.js';
-export * from './svelte-autofixer.js';
-export * from './playground-link.js';
+export { get_documentation } from './get-documentation.js';
+export { list_sections } from './list-sections.js';
+export { svelte_autofixer } from './svelte-autofixer.js';
+export { playground_link } from './playground-link.js';

packages/mcp-server/src/mcp/handlers/tools/list-sections.ts

@@ -2,6 +2,13 @@ import type { SvelteMcp } from '../../index.js';
 import { format_sections_list } from '../../utils.js';
 import { SECTIONS_LIST_INTRO, SECTIONS_LIST_OUTRO } from './prompts.js';
 import { icons } from '../../icons/index.js';
+import { tool } from 'tmcp/utils';
+
+export async function list_sections_handler() {
+	const formatted_sections = await format_sections_list();
+
+	return `${SECTIONS_LIST_INTRO}\n\n${formatted_sections}\n\n${SECTIONS_LIST_OUTRO}`;
+}
 
 export function list_sections(server: SvelteMcp) {
 	server.tool(
@@ -9,19 +16,32 @@ export function list_sections(server: SvelteMcp) {
 			name: 'list-sections',
 			description:
 				'Lists all available Svelte 5 and SvelteKit documentation sections in a structured format. Each section includes a "use_cases" field that describes WHEN this documentation would be useful. You should carefully analyze the use_cases field to determine which sections are relevant for the user\'s query. The use_cases contain comma-separated keywords describing project types (e.g., "e-commerce", "blog"), features (e.g., "authentication", "forms"), components (e.g., "slider", "modal"), development stages (e.g., "deployment", "testing"), or "always" for fundamental concepts. Match these use_cases against the user\'s intent - for example, if building an e-commerce site, fetch sections with use_cases containing "e-commerce", "product listings", "shopping cart", etc. If building a slider, look for "slider", "carousel", "animation", etc. Returns sections as "* title: [section_title], use_cases: [use_cases], path: [file_path]". Always run list-sections FIRST for any Svelte query, then analyze ALL use_cases to identify relevant sections, and finally use get_documentation to fetch ALL relevant sections at once.',
+			annotations: {
+				title: 'List Sections',
+				destructiveHint: false,
+				readOnlyHint: true,
+				openWorldHint: false,
+			},
 			icons,
 		},
 		async () => {
-			const formatted_sections = await format_sections_list();
-
-			return {
-				content: [
-					{
-						type: 'text',
-						text: `${SECTIONS_LIST_INTRO}\n\n${formatted_sections}\n\n${SECTIONS_LIST_OUTRO}`,
-					},
-				],
-			};
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(server.ctx.sessionId, 'list-sections');
+			}
+			try {
+				const content = await list_sections_handler();
+				return tool.text(content);
+			} catch (e) {
+				const error = e as Error;
+				if (server.ctx.sessionId && server.ctx.custom?.track) {
+					await server.ctx.custom?.track?.(
+						server.ctx.sessionId,
+						'list-sections-error',
+						error.message,
+					);
+				}
+				return tool.error(error.message);
+			}
 		},
 	);
 }

packages/mcp-server/src/mcp/handlers/tools/playground-link.test.ts

@@ -0,0 +1,102 @@
+import { InMemoryTransport } from '@tmcp/transport-in-memory';
+import { beforeEach, describe, expect, it } from 'vitest';
+import { server } from '../../index.js';
+
+const transport = new InMemoryTransport(server);
+
+let session: ReturnType<typeof transport.session>;
+
+describe('playground-link tool', () => {
+	beforeEach(async () => {
+		session = transport.session();
+		await session.initialize(
+			'2025-06-18',
+			{},
+			{
+				name: 'test-client',
+				version: '1.0.0',
+			},
+		);
+	});
+
+	it('should create a playground link if App.svelte is present', async () => {
+		const result = await session.callTool<{ url: string }>('playground-link', {
+			name: 'My Playground',
+			tailwind: false,
+			files: {
+				'App.svelte': `Hi there!`,
+			},
+		});
+		expect(result.structuredContent).toBeDefined();
+		expect(result.structuredContent?.url).toBeDefined();
+		// Verify URL structure rather than exact match (gzip compression can vary by platform)
+		expect(result.structuredContent?.url).toMatch(/^https:\/\/svelte\.dev\/playground#H4sIA/);
+		expect(result.structuredContent?.url).toContain('svelte.dev/playground');
+	});
+
+	it('should have a content with the stringified version of structured content and an ui resource', async () => {
+		const result = await session.callTool<{ url: string }>('playground-link', {
+			name: 'My Playground',
+			tailwind: false,
+			files: {
+				'App.svelte': `Hi there!`,
+			},
+		});
+		expect(result.structuredContent).toBeDefined();
+		expect(result.content).toStrictEqual(
+			expect.arrayContaining([
+				expect.objectContaining({
+					type: 'text',
+					text: JSON.stringify(result.structuredContent),
+				}),
+			]),
+		);
+		// Verify resource structure without exact URL match (gzip compression can vary by platform)
+		expect(result.content).toStrictEqual(
+			expect.arrayContaining([
+				expect.objectContaining({
+					type: 'resource',
+					resource: expect.objectContaining({
+						uri: 'ui://svelte/playground-link',
+						mimeType: 'text/html;profile=mcp-app',
+						_meta: { 'mcpui.dev/ui-preferred-frame-size': ['100%', '1200px'] },
+						text: expect.stringMatching(/^https:\/\/svelte\.dev\/playground\/embed#H4sIA/),
+					}),
+				}),
+			]),
+		);
+	});
+
+	it('should have tool _meta with resource URI for MCP Apps hosts', async () => {
+		const tools = await session.listTools();
+		const playground_tool = tools.tools.find((t) => t.name === 'playground-link');
+		expect(playground_tool).toBeDefined();
+		expect(playground_tool?._meta).toStrictEqual({
+			ui: { resourceUri: 'ui://svelte/playground-link' },
+		});
+	});
+
+	it('should expose a resource for MCP Apps hosts', async () => {
+		const resources = await session.listResources();
+		const playground_resource = resources.resources.find(
+			(r) => r.uri === 'ui://svelte/playground-link',
+		);
+		expect(playground_resource).toBeDefined();
+		expect(playground_resource?.name).toBe('playground-link-ui');
+	});
+
+	it('should not create a playground link if App.svelte is missing', async () => {
+		const result = await session.callTool<{ url: string }>('playground-link', {
+			name: 'My Playground',
+			tailwind: false,
+			files: {
+				'Something.svelte': `Hi there!`,
+			},
+		});
+		expect(result.isError).toBe(true);
+		expect(result.content?.[0]).toStrictEqual({
+			type: 'text',
+			text: 'The files must contain an App.svelte file as the entry point',
+		});
+	});
+});

packages/mcp-server/src/mcp/handlers/tools/playground-link.ts

@@ -1,7 +1,8 @@
-import type { SvelteMcp } from '../../index.js';
+import { createUIResource } from '@mcp-ui/server';
+import { tool } from 'tmcp/utils';
 import * as v from 'valibot';
 import { icons } from '../../icons/index.js';
-import { createUIResource } from '@mcp-ui/server';
+import type { SvelteMcp } from '../../index.js';
 
 async function compress_and_encode_text(input: string) {
 	const reader = new Blob([input]).stream().pipeThrough(new CompressionStream('gzip')).getReader();
@@ -29,101 +30,232 @@ type File = {
 	text: boolean;
 };
 
+const playground_link_schema = v.object({
+	name: v.pipe(
+		v.string(),
+		v.description('The name of the Playground, it should reflect the user task'),
+	),
+	tailwind: v.pipe(
+		v.boolean(),
+		v.description(
+			"If the code requires Tailwind CSS to work...only send true if it it's using tailwind classes in the code",
+		),
+	),
+	files: v.pipe(
+		v.record(v.string(), v.string()),
+		v.description(
+			"An object where all the keys are the filenames (with extensions) and the values are the file content. For example: { 'Component.svelte': '<script>...</script>', 'utils.js': 'export function ...' }. The playground accept multiple files so if are importing from other files just include them all at the root level.",
+		),
+	),
+});
+
+const playground_link_output_schema = v.object({
+	url: v.string(),
+});
+
+export async function playground_link_handler({
+	files,
+	name,
+	tailwind,
+}: v.InferInput<typeof playground_link_schema>) {
+	const playground_base = new URL('https://svelte.dev/playground');
+	const playground_files: File[] = [];
+
+	let has_app_svelte = false;
+
+	for (const [filename, contents] of Object.entries(files)) {
+		if (filename === 'App.svelte') has_app_svelte = true;
+		playground_files.push({
+			type: 'file',
+			name: filename,
+			basename: filename.replace(/^.*[\\/]/, ''),
+			contents,
+			text: true,
+		});
+	}
+
+	if (!has_app_svelte) {
+		throw new Error('The files must contain an App.svelte file as the entry point');
+	}
+
+	const playground_config = {
+		name,
+		tailwind: tailwind ?? false,
+		files: playground_files,
+	};
+
+	playground_base.hash = await compress_and_encode_text(JSON.stringify(playground_config));
+
+	const url = playground_base.toString();
+
+	// use the embed path to have a cleaner UI for mcp-ui
+	playground_base.pathname = '/playground/embed';
+
+	return {
+		url,
+		iframe_url: playground_base.toString(),
+	};
+}
+
+// Create the UI resource for MCP Apps hosts (with adapter)
+// This will be registered as a resource that MCP Apps hosts can fetch
+const playground_ui_resource = createUIResource({
+	uri: 'ui://svelte/playground-link',
+	encoding: 'text',
+	resourceProps: {
+		_meta: {
+			ui: {
+				csp: {
+					connectDomains: ['https://svelte.dev'],
+					resourceDomains: ['https://svelte.dev'],
+					frameDomains: ['https://svelte.dev'],
+					baseUriDomains: ['https://svelte.dev'],
+				},
+			},
+		},
+	},
+	content: {
+		type: 'rawHtml',
+		// This is a placeholder HTML - the actual iframe URL will be set per-request
+		// MCP Apps hosts receive the tool input/output via postMessage
+		htmlString: `<!DOCTYPE html>
+<html>
+<head>
+	<meta charset="utf-8">
+	<meta name="viewport" content="width=device-width, initial-scale=1">
+	<style>
+		* { margin: 0; padding: 0; box-sizing: border-box; }
+		html, body { width: 100%; height: 100%; }
+		iframe { width: 100%; height: 100%; border: none; display: none; }
+		.loading { display: flex; align-items: center; justify-content: center; height: 100%; font-family: system-ui, sans-serif; color: #666; }
+	</style>
+</head>
+<body>
+	<div class="loading" id="loading">Loading playground...</div>
+	<iframe id="playground" allow="clipboard-write"></iframe>
+	<script>
+		function size_changed() {
+			const width = document.body.scrollWidth;
+			window.parent.postMessage({
+				jsonrpc: '2.0',
+				method: 'ui/notifications/size-changed',
+				params: {
+					width,
+					height: 800
+					}
+				}, '*');
+		}
+		// Signal that the widget is ready
+		window.parent.postMessage({ type: 'ui-lifecycle-iframe-ready' }, '*');
+		
+		// Listen for render data from the adapter (for MCP Apps hosts)
+		window.addEventListener('message', (event) => {
+			if (event.data.type === 'ui-lifecycle-iframe-render-data') {
+				const renderData = event.data.payload.renderData || {};
+				const toolOutput = renderData.toolOutput;
+				
+				// The tool output contains the iframe URL
+				if (toolOutput && toolOutput.structuredContent && toolOutput.structuredContent.url) {
+					const iframe = document.getElementById('playground');
+					const loading = document.getElementById('loading');
+					// Convert the URL to embed URL
+					const embedUrl = toolOutput.structuredContent.url.replace('/playground#', '/playground/embed#');
+					iframe.src = embedUrl;
+					iframe.style.display = 'block';
+					iframe.addEventListener("load", () => {
+						size_changed();
+					});
+					loading.style.display = 'none';
+				}
+			}
+		});
+	</script>
+</body>
+</html>`,
+	},
+	uiMetadata: {
+		'preferred-frame-size': ['100%', '1200px'],
+	},
+	adapters: {
+		mcpApps: { enabled: true },
+	},
+});
+
 export function playground_link(server: SvelteMcp) {
+	// Register the UI resource so MCP Apps hosts can fetch it
+	server.resource(
+		{
+			name: 'playground-link-ui',
+			description: 'UI resource for the Svelte Playground widget',
+			uri: playground_ui_resource.resource.uri,
+			icons,
+		},
+		() => {
+			return {
+				contents: [playground_ui_resource.resource],
+			};
+		},
+	);
+
 	server.tool(
 		{
 			name: 'playground-link',
 			description:
 				'Generates a Playground link given a Svelte code snippet. Once you have the final version of the code you want to send to the user, ALWAYS ask the user if it wants a playground link to allow it to quickly check the code in the playground before calling this tool. NEVER use this tool if you have written the component to a file in the user project. The playground accept multiple files so if are importing from other files just include them all at the root level.',
-			schema: v.object({
-				name: v.pipe(
-					v.string(),
-					v.description('The name of the Playground, it should reflect the user task'),
-				),
-				tailwind: v.pipe(
-					v.boolean(),
-					v.description(
-						"If the code requires Tailwind CSS to work...only send true if it it's using tailwind classes in the code",
-					),
-				),
-				files: v.pipe(
-					v.record(v.string(), v.string()),
-					v.description(
-						"An object where all the keys are the filenames (with extensions) and the values are the file content. For example: { 'Component.svelte': '<script>...</script>', 'utils.js': 'export function ...' }. The playground accept multiple files so if are importing from other files just include them all at the root level.",
-					),
-				),
-			}),
-			outputSchema: v.object({
-				url: v.string(),
-			}),
+			schema: playground_link_schema,
+			outputSchema: playground_link_output_schema,
+			annotations: {
+				title: 'Playground Link',
+				destructiveHint: false,
+				readOnlyHint: true,
+				openWorldHint: false,
+			},
 			icons,
+			// For MCP Apps hosts - points to the registered resource
+			_meta: {
+				ui: {
+					resourceUri: playground_ui_resource.resource.uri,
+				},
+			},
 		},
 		async ({ files, name, tailwind }) => {
-			const playground_base = new URL('https://svelte.dev/playground');
-			const playground_files: File[] = [];
-
-			let has_app_svelte = false;
-
-			for (const [filename, contents] of Object.entries(files)) {
-				if (filename === 'App.svelte') has_app_svelte = true;
-				playground_files.push({
-					type: 'file',
-					name: filename,
-					basename: filename.replace(/^.*[\\/]/, ''),
-					contents,
-					text: true,
-				});
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(server.ctx.sessionId, 'playground-link');
 			}
-
-			if (!has_app_svelte) {
+			try {
+				const result = await playground_link_handler({ files, name, tailwind });
 				return {
-					isError: true,
 					content: [
 						{
 							type: 'text',
-							text: JSON.stringify({
-								error: 'The files must contain an App.svelte file as the entry point',
-							}),
+							text: JSON.stringify({ url: result.url }),
 						},
+						// Embedded resource for MCP-UI hosts (no adapter, uses externalUrl)
+						createUIResource({
+							uri: 'ui://svelte/playground-link',
+							content: {
+								type: 'externalUrl',
+								iframeUrl: result.iframe_url,
+							},
+							uiMetadata: {
+								'preferred-frame-size': ['100%', '1200px'],
+							},
+							encoding: 'text',
+						}),
 					],
+					structuredContent: { url: result.url },
 				};
+			} catch (e) {
+				const error = e as Error;
+				if (server.ctx.sessionId && server.ctx.custom?.track) {
+					await server.ctx.custom?.track?.(
+						server.ctx.sessionId,
+						'playground-link-error',
+						error.message,
+					);
+				}
+				return tool.error(error.message);
 			}
-
-			const playground_config = {
-				name,
-				tailwind: tailwind ?? false,
-				files: playground_files,
-			};
-
-			playground_base.hash = await compress_and_encode_text(JSON.stringify(playground_config));
-
-			const content = {
-				url: playground_base.toString(),
-			};
-
-			// use the embed path to have a cleaner UI for mcp-ui
-			playground_base.pathname = '/playground/embed';
-
-			return {
-				content: [
-					{
-						type: 'text',
-						text: JSON.stringify(content),
-					},
-					createUIResource({
-						uri: 'ui://svelte/playground-link',
-						content: {
-							type: 'externalUrl',
-							iframeUrl: playground_base.toString(),
-						},
-						uiMetadata: {
-							'preferred-frame-size': ['100%', '1200px'],
-						},
-						encoding: 'text',
-					}),
-				],
-				structuredContent: content,
-			};
 		},
 	);
 }

packages/mcp-server/src/mcp/handlers/tools/svelte-autofixer.test.ts

@@ -1,58 +1,46 @@
+/* eslint-disable @typescript-eslint/no-explicit-any */
+import { InMemoryTransport } from '@tmcp/transport-in-memory';
 import { beforeEach, describe, expect, it } from 'vitest';
 import { server } from '../../index.js';
 
-/**
- * Small utility to create a JSON-RPC request without having to always specify as const
- */
-function request<const T>(request: T) {
-	return request;
-}
+const transport = new InMemoryTransport(server);
 
-async function autofixer_tool_call(code: string, is_error = false, desired_svelte_version = 5) {
-	const result = await server.receive({
-		jsonrpc: '2.0',
-		id: 2,
-		method: 'tools/call',
-		params: {
-			name: 'svelte-autofixer',
-			arguments: {
-				code,
-				desired_svelte_version,
-				filename: 'App.svelte',
-			},
-		},
+let session: ReturnType<typeof transport.session>;
+
+async function autofixer_tool_call(
+	code: string,
+	is_error = false,
+	desired_svelte_version = 5,
+	async = false,
+) {
+	const result = await session.callTool('svelte-autofixer', {
+		code,
+		desired_svelte_version,
+		filename: 'App.svelte',
+		async,
 	});
 
 	expect(result).toBeDefined();
-	expect(result.result).toBeDefined();
 	if (is_error) {
-		return result.result;
+		return result as any;
 	}
-	expect(result.result.structuredContent).toBeDefined();
-	return result.result.structuredContent;
+	expect(result.structuredContent).toBeDefined();
+	return result.structuredContent as any;
 }
 
 describe('svelte-autofixer tool', () => {
 	beforeEach(async () => {
-		const initialize_request = request({
-			jsonrpc: '2.0',
-			id: 1,
-			method: 'initialize',
-			params: {
-				protocolVersion: '2025-06-18',
-				capabilities: {
-					roots: { listChanged: true },
-				},
-				clientInfo: {
-					name: 'test-client',
-					version: '1.0.0',
-				},
-			},
-		});
+		session = transport.session();
 
-		await server.receive(initialize_request, {
-			sessionId: 'svelte-autofixer-session',
-		});
+		session = transport.session();
+		await session.initialize(
+			'2025-06-18',
+			{},
+			{
+				name: 'test-client',
+				version: '1.0.0',
+			},
+		);
 	});
 
 	it('should add suggestions for js parse errors', async () => {
@@ -65,6 +53,73 @@ describe('svelte-autofixer tool', () => {
 		);
 	});
 
+	it('should add suggestions for snippets declared in script tag', async () => {
+		const content = await autofixer_tool_call(`<script>
+			{#snippet my_snippet()}
+				some content
+			{/snippet}
+		</script>`);
+		expect(content.issues.length).toBeGreaterThan(0);
+		expect(content.suggestions).toContain(
+			"The code can't be compiled because a Javascript parse error. The error suggests you have a `{#snippet ...}` block inside the `<script>` tag. Snippets are template syntax and should be declared in the markup section of the component, not in the script. Move the snippet outside of the `<script>` tag. Snippets declared in the markup can also be accessed in the script tag in case you need them.",
+		);
+	});
+
+	it('should error out if async is true with a version less than 5', async () => {
+		const content = await autofixer_tool_call(
+			`<script>
+			$state count = 0;
+		</script>`,
+			true,
+			4,
+			true,
+		);
+		expect(content.isError).toBeTruthy();
+		expect(content.content[0]).toBeDefined();
+		expect(content.content[0].text).toBe(
+			'The async option can only be used with Svelte version 5 or higher.',
+		);
+	});
+
+	it('should not add suggestion/issues if async is true and await is used in the template/derived', async () => {
+		const content = await autofixer_tool_call(
+			`<script>
+			import { slow_double } from './utils.js';
+			let count = $state(0);
+			let double = $derived(await slow_double(count));
+		</script>
+		
+		{double}
+		{await slow_double(count)}`,
+			false,
+			5,
+			true,
+		);
+		expect(content.issues).toHaveLength(0);
+		expect(content.suggestions).toHaveLength(0);
+		expect(content.require_another_tool_call_after_fixing).toBeFalsy();
+	});
+
+	it('should add suggestion/issues if async is false and await is used in the template/derived', async () => {
+		const content = await autofixer_tool_call(
+			`<script>
+			import { slow_double } from './utils.js';
+			let count = $state(0);
+			let double = $derived(await slow_double(count));
+		</script>
+		
+		{double}
+		{await slow_double(count)}`,
+			false,
+			5,
+		);
+		expect(content.issues.length).toBeGreaterThanOrEqual(1);
+		expect(content.issues).toEqual(
+			expect.arrayContaining([expect.stringContaining('experimental_async')]),
+		);
+		expect(content.require_another_tool_call_after_fixing).toBeTruthy();
+	});
+
 	it('should add suggestions for css invalid identifier', async () => {
 		const content = await autofixer_tool_call(`<script>
 				let my_color = $state('red');

packages/mcp-server/src/mcp/handlers/tools/svelte-autofixer.ts

@@ -5,6 +5,103 @@ import { add_compile_issues } from '../../autofixers/add-compile-issues.js';
 import { add_eslint_issues } from '../../autofixers/add-eslint-issues.js';
 import { add_autofixers_issues } from '../../autofixers/add-autofixers-issues.js';
 import { icons } from '../../icons/index.js';
+import { tool } from 'tmcp/utils';
+
+const autofixer_schema = v.object({
+	code: v.string(),
+	desired_svelte_version: v.pipe(
+		v.union([v.string(), v.number()]),
+		v.description(
+			'The desired svelte version...if possible read this from the package.json of the user project, otherwise use some hint from the wording (if the user asks for runes it wants version 5). Default to 5 in case of doubt.',
+		),
+	),
+	async: v.pipe(
+		v.optional(v.boolean()),
+		v.description(
+			'If true the code is an async component/module and might use await in the markup or top-level awaits in the script tag. If possible check the svelte.config.js/svelte.config.ts to check if the option is enabled otherwise asks the user if they prefer using it or not. You can only use this option if the version is 5.',
+		),
+	),
+	filename: v.pipe(
+		v.optional(v.string()),
+		v.description(
+			'The filename of the component if available, it MUST be only the Component name with .svelte or .svelte.ts extension and not the entire path.',
+		),
+	),
+});
+
+const autofixer_output_schema = v.object({
+	issues: v.array(v.string()),
+	suggestions: v.array(v.string()),
+	require_another_tool_call_after_fixing: v.boolean(),
+});
+
+export async function svelte_autofixer_handler({
+	code,
+	desired_svelte_version: desired_svelte_version_unchecked,
+	async,
+	filename: filename_or_path,
+}: v.InferInput<typeof autofixer_schema>) {
+	// we validate manually because some clients don't support union in the input schema (looking at you cursor)
+	const parsed_version = v.safeParse(
+		v.union([v.literal(4), v.literal(5), v.literal('4'), v.literal('5')]),
+		desired_svelte_version_unchecked,
+	);
+	if (parsed_version.success === false) {
+		throw new Error(
+			`The desired_svelte_version MUST be either 4 or 5 but received "${desired_svelte_version_unchecked}"`,
+		);
+	}
+
+	const desired_svelte_version = parsed_version.output;
+
+	if (async && +desired_svelte_version < 5) {
+		throw new Error('The async option can only be used with Svelte version 5 or higher.');
+	}
+
+	const content: {
+		issues: string[];
+		suggestions: string[];
+		require_another_tool_call_after_fixing: boolean;
+	} = { issues: [], suggestions: [], require_another_tool_call_after_fixing: false };
+	try {
+		// just in case the LLM sends a full path we extract the filename...it's not really needed
+		// but it's nice to have a filename in the errors
+
+		const filename = filename_or_path ? basename(filename_or_path) : 'Component.svelte';
+
+		add_compile_issues(content, code, +desired_svelte_version, filename, async);
+
+		add_autofixers_issues(content, code, +desired_svelte_version, filename, async);
+
+		await add_eslint_issues(content, code, +desired_svelte_version, filename, async);
+	} catch (e: unknown) {
+		const error = e as Error & { start?: { line: number; column: number }; frame?: string };
+		content.issues.push(
+			`${error.message} at line ${error.start?.line}, column ${error.start?.column}`,
+		);
+		if (error.message.includes('js_parse_error')) {
+			// Check if the error frame contains template syntax that was incorrectly placed in the script tag
+			if (error.frame?.includes('{#snippet')) {
+				content.suggestions.push(
+					"The code can't be compiled because a Javascript parse error. The error suggests you have a `{#snippet ...}` block inside the `<script>` tag. Snippets are template syntax and should be declared in the markup section of the component, not in the script. Move the snippet outside of the `<script>` tag. Snippets declared in the markup can also be accessed in the script tag in case you need them.",
+				);
+			} else {
+				content.suggestions.push(
+					"The code can't be compiled because a Javascript parse error. In case you are using runes like this `$state variable_name = 3;` or `$derived variable_name = 3 * count` that's not how runes are used. You need to use them as function calls without importing them: `const variable_name = $state(3)` and `const variable_name = $derived(3 * count)`.",
+				);
+			}
+		} else if (error.message.includes('css_expected_identifier')) {
+			content.suggestions.push(
+				"The code can't be compiled because a valid CSS identifier is expected. This sometimes means you are trying to use a variable in CSS like this: `color: {my_color}` but Svelte doesn't support that. You can use inline CSS variables for that `<div style:--color={my_color}></div>` and then use the variable as usual in CSS with `color: var(--color)`.",
+			);
+		}
+	}
+
+	if (content.issues.length > 0 || content.suggestions.length > 0) {
+		content.require_another_tool_call_after_fixing = true;
+	}
+	return content;
+}
 
 export function svelte_autofixer(server: SvelteMcp) {
 	server.tool(
@@ -13,26 +110,8 @@ export function svelte_autofixer(server: SvelteMcp) {
 			title: 'Svelte Autofixer',
 			description:
 				'Given a svelte component or module returns a list of suggestions to fix any issues it has. This tool MUST be used whenever the user is asking to write svelte code before sending the code back to the user',
-			schema: v.object({
-				code: v.string(),
-				desired_svelte_version: v.pipe(
-					v.union([v.string(), v.number()]),
-					v.description(
-						'The desired svelte version...if possible read this from the package.json of the user project, otherwise use some hint from the wording (if the user asks for runes it wants version 5). Default to 5 in case of doubt.',
-					),
-				),
-				filename: v.pipe(
-					v.optional(v.string()),
-					v.description(
-						'The filename of the component if available, it MUST be only the Component name with .svelte or .svelte.ts extension and not the entire path.',
-					),
-				),
-			}),
-			outputSchema: v.object({
-				issues: v.array(v.string()),
-				suggestions: v.array(v.string()),
-				require_another_tool_call_after_fixing: v.boolean(),
-			}),
+			schema: autofixer_schema,
+			outputSchema: autofixer_output_schema,
 			annotations: {
 				title: 'Svelte Autofixer',
 				destructiveHint: false,
@@ -45,71 +124,30 @@ export function svelte_autofixer(server: SvelteMcp) {
 			code,
 			filename: filename_or_path,
 			desired_svelte_version: desired_svelte_version_unchecked,
+			async,
 		}) => {
-			// we validate manually because some clients don't support union in the input schema (looking at you cursor)
-			const parsed_version = v.safeParse(
-				v.union([v.literal(4), v.literal(5), v.literal('4'), v.literal('5')]),
-				desired_svelte_version_unchecked,
-			);
-			if (parsed_version.success === false) {
-				return {
-					isError: true,
-					content: [
-						{
-							type: 'text',
-							text: `The desired_svelte_version MUST be either 4 or 5 but received "${desired_svelte_version_unchecked}"`,
-						},
-					],
-				};
+			if (server.ctx.sessionId && server.ctx.custom?.track) {
+				await server.ctx.custom?.track?.(server.ctx.sessionId, 'svelte-autofixer');
 			}
-
-			const desired_svelte_version = parsed_version.output;
-
-			const content: {
-				issues: string[];
-				suggestions: string[];
-				require_another_tool_call_after_fixing: boolean;
-			} = { issues: [], suggestions: [], require_another_tool_call_after_fixing: false };
 			try {
-				// just in case the LLM sends a full path we extract the filename...it's not really needed
-				// but it's nice to have a filename in the errors
-
-				const filename = filename_or_path ? basename(filename_or_path) : 'Component.svelte';
-
-				add_compile_issues(content, code, +desired_svelte_version, filename);
-
-				add_autofixers_issues(content, code, +desired_svelte_version, filename);
-
-				await add_eslint_issues(content, code, +desired_svelte_version, filename);
-			} catch (e: unknown) {
-				const error = e as Error & { start?: { line: number; column: number } };
-				content.issues.push(
-					`${error.message} at line ${error.start?.line}, column ${error.start?.column}`,
-				);
-				if (error.message.includes('js_parse_error')) {
-					content.suggestions.push(
-						"The code can't be compiled because a Javascript parse error. In case you are using runes like this `$state variable_name = 3;` or `$derived variable_name = 3 * count` that's not how runes are used. You need to use them as function calls without importing them: `const variable_name = $state(3)` and `const variable_name = $derived(3 * count)`.",
-					);
-				} else if (error.message.includes('css_expected_identifier')) {
-					content.suggestions.push(
-						"The code can't be compiled because a valid CSS identifier is expected. This sometimes means you are trying to use a variable in CSS like this: `color: {my_color}` but Svelte doesn't support that. You can use inline CSS variables for that `<div style:--color={my_color}></div>` and then use the variable as usual in CSS with `color: var(--color)`.",
+				const content = await svelte_autofixer_handler({
+					code,
+					desired_svelte_version: desired_svelte_version_unchecked,
+					async,
+					filename: filename_or_path,
+				});
+				return tool.structured(content);
+			} catch (e) {
+				const error = e as Error;
+				if (server.ctx.sessionId && server.ctx.custom?.track) {
+					await server.ctx.custom?.track?.(
+						server.ctx.sessionId,
+						'svelte-autofixer-error',
+						error.message,
 					);
 				}
+				return tool.error(error.message);
 			}
-
-			if (content.issues.length > 0 || content.suggestions.length > 0) {
-				content.require_another_tool_call_after_fixing = true;
-			}
-
-			return {
-				content: [
-					{
-						type: 'text',
-						text: JSON.stringify(content),
-					},
-				],
-				structuredContent: content,
-			};
 		},
 	);
 }

packages/mcp-server/src/mcp/index.ts

@@ -1,8 +1,6 @@
 import { ValibotJsonSchemaAdapter } from '@tmcp/adapter-valibot';
 import { McpServer } from 'tmcp';
 import { setup_prompts, setup_resources, setup_tools } from './handlers/index.js';
-import type { LibSQLDatabase } from 'drizzle-orm/libsql';
-import type { Schema } from '@sveltejs/mcp-schema';
 import { icons } from './icons/index.js';
 
 export const server = new McpServer(
@@ -24,10 +22,17 @@ export const server = new McpServer(
 		instructions:
 			'This is the official Svelte MCP server. It MUST be used whenever svelte development is involved. It can provide official documentation, code examples and correct your code. After you correct the component call this tool again to confirm all the issues are fixed.',
 	},
-).withContext<{ db: LibSQLDatabase<Schema> }>();
+).withContext<{
+	track?: (sessionId: string, event: string, extra?: string) => Promise<void>;
+}>();
 
 export type SvelteMcp = typeof server;
 
 setup_tools(server);
 setup_resources(server);
 setup_prompts(server);
+
+server.on('initialize', async ({ clientInfo: client_info }) => {
+	if (!server.ctx.custom?.track || !server.ctx.sessionId) return;
+	server.ctx.custom.track(server.ctx.sessionId, 'initialize', client_info.name);
+});

packages/mcp-server/src/parse/parse.ts

@@ -69,6 +69,8 @@ export function parse(code: string, file_path: string) {
 			return this.all_references.find((r) => r.identifier === id);
 		},
 		is_rune(call: CallExpression, rune?: (typeof runes)[number][]) {
+			// it should always be a call expression but we check just in case because sometimes it's any and TS doesn't complain
+			if (call.type !== 'CallExpression') return false;
 			if (call.callee.type !== 'Identifier' && call.callee.type !== 'MemberExpression')
 				return false;
 			const id = call.callee.type === 'Identifier' ? call.callee : call.callee.object;

packages/mcp-stdio/CHANGELOG.md

@@ -1,111 +1,187 @@
 # @sveltejs/mcp
 
+## 0.1.22
+
+### Patch Changes
+
+- fix: broaden checks for imported runes because LLMs are unhinged ([#185](https://github.com/sveltejs/ai-tools/pull/185))
+
+## 0.1.21
+
+### Patch Changes
+
+- feat: display similar result & error at the end ([#161](https://github.com/sveltejs/ai-tools/pull/161))
+
+## 0.1.20
+
+### Patch Changes
+
+- fix: turn off no-inspect in eslint for mcp ([`2245cb2`](https://github.com/sveltejs/ai-tools/commit/2245cb2dc9e2d217869b6a800795ce59ffb40c51))
+
+## 0.1.19
+
+### Patch Changes
+
+- chore: update svelte ([`7447744`](https://github.com/sveltejs/ai-tools/commit/74477448cea44ec21684ea4d39f2c5c7133b5150))
+
+## 0.1.18
+
+### Patch Changes
+
+- feat: expose playground link as MCP App ([#138](https://github.com/sveltejs/ai-tools/pull/138))
+
+## 0.1.17
+
+### Patch Changes
+
+- fix: add suggestion for snippets declared in script tag ([#132](https://github.com/sveltejs/ai-tools/pull/132))
+
+## 0.1.16
+
+### Patch Changes
+
+- feat: expose tools as JS api + cli ([#128](https://github.com/sveltejs/ai-tools/pull/128))
+
+## 0.1.15
+
+### Patch Changes
+
+- fix: server.json version + update publisher ([`9dfb4de`](https://github.com/sveltejs/ai-tools/commit/9dfb4dedb42837c40c4e660f0f816d7cf9081fc4))
+
+## 0.1.14
+
+### Patch Changes
+
+- fix: improve prompt to reduce token usage ([#124](https://github.com/sveltejs/ai-tools/pull/124))
+
+## 0.1.13
+
+### Patch Changes
+
+- fix: revert name change and add title ([`98efa1e`](https://github.com/sveltejs/ai-tools/commit/98efa1e09ebcca7827b10dc6bc8e1699fc1e5171))
+
+## 0.1.12
+
+### Patch Changes
+
+- fix: update server name on mcp registry ([`60297b3`](https://github.com/sveltejs/ai-tools/commit/60297b3c49bf110b48908e61b5d5d902ea1bdf39))
+
+- chore: update tmcp ([#99](https://github.com/sveltejs/ai-tools/pull/99))
+
+## 0.1.11
+
+### Patch Changes
+
+- fix: add `async` parameter to `svelte-autofixer` ([#94](https://github.com/sveltejs/ai-tools/pull/94))
+
+- fix: install latest eslint svelte packages to support `$state.eager` ([`f6ce89f`](https://github.com/sveltejs/ai-tools/commit/f6ce89ff34faabc3d746a350ea347298ecfed2ec))
+
 ## 0.1.10
 
 ### Patch Changes
 
-- fix: add icons to `server.json` ([`02c951b`](https://github.com/sveltejs/mcp/commit/02c951baa86ac8103ffc158a202c06cfe6b15c01))
+- fix: add icons to `server.json` ([`02c951b`](https://github.com/sveltejs/ai-tools/commit/02c951baa86ac8103ffc158a202c06cfe6b15c01))
 
-- fix: add `preferred-frame-size` to UI resource ([`3fabcc0`](https://github.com/sveltejs/mcp/commit/3fabcc0f9bfee916c0deb9c2ffa931ed2168af2d))
+- fix: add `preferred-frame-size` to UI resource ([`3fabcc0`](https://github.com/sveltejs/ai-tools/commit/3fabcc0f9bfee916c0deb9c2ffa931ed2168af2d))
 
-- feat: support: `$state.eager` ([#90](https://github.com/sveltejs/mcp/pull/90))
+- feat: support: `$state.eager` ([#90](https://github.com/sveltejs/ai-tools/pull/90))
 
 ## 0.1.9
 
 ### Patch Changes
 
-- feat: return `mcp-ui` resource from `playground-link` ([#84](https://github.com/sveltejs/mcp/pull/84))
+- feat: return `mcp-ui` resource from `playground-link` ([#84](https://github.com/sveltejs/ai-tools/pull/84))
 
-- feat: suggest against js variables in css ([#78](https://github.com/sveltejs/mcp/pull/78))
+- feat: suggest against js variables in css ([#78](https://github.com/sveltejs/ai-tools/pull/78))
 
 ## 0.1.8
 
 ### Patch Changes
 
-- fix: upgrade registry publisher cli ([`5fa2baa`](https://github.com/sveltejs/mcp/commit/5fa2baa27009f01e0e4e91cee7984b81a81c1c29))
+- fix: upgrade registry publisher cli ([`5fa2baa`](https://github.com/sveltejs/ai-tools/commit/5fa2baa27009f01e0e4e91cee7984b81a81c1c29))
 
 ## 0.1.7
 
 ### Patch Changes
 
-- fix: use correct server schema version ([`579be87`](https://github.com/sveltejs/mcp/commit/579be877fa9f87f7f173450ca5bc918824d68282))
+- fix: use correct server schema version ([`579be87`](https://github.com/sveltejs/ai-tools/commit/579be877fa9f87f7f173450ca5bc918824d68282))
 
 ## 0.1.6
 
 ### Patch Changes
 
-- fix: prevent `imported_runes` suggestion from being added for libs that are not svelte ([`87af64f`](https://github.com/sveltejs/mcp/commit/87af64f4bc6d07b75640eb987a33655654363997))
+- fix: prevent `imported_runes` suggestion from being added for libs that are not svelte ([`87af64f`](https://github.com/sveltejs/ai-tools/commit/87af64f4bc6d07b75640eb987a33655654363997))
 
-- feat: add svelte icon and website url for mcp server ([#75](https://github.com/sveltejs/mcp/pull/75))
+- feat: add svelte icon and website url for mcp server ([#75](https://github.com/sveltejs/ai-tools/pull/75))
 
-- fix: use `data:` uri for local icon & add icons to tools + resources + prompts ([`cf62286`](https://github.com/sveltejs/mcp/commit/cf622869129382a97ad059bb1389f115907adc8e))
+- fix: use `data:` uri for local icon & add icons to tools + resources + prompts ([`cf62286`](https://github.com/sveltejs/ai-tools/commit/cf622869129382a97ad059bb1389f115907adc8e))
 
 ## 0.1.5
 
 ### Patch Changes
 
-- fix: widen `desired_svelte_version` validation to accommodate some clients ([`3b301d7`](https://github.com/sveltejs/mcp/commit/3b301d7d9c2f49758023408f505bc4ca79caaff4))
+- fix: widen `desired_svelte_version` validation to accommodate some clients ([`3b301d7`](https://github.com/sveltejs/ai-tools/commit/3b301d7d9c2f49758023408f505bc4ca79caaff4))
 
-- fix: minor tweaks to the prompt to allow for automatic sync ([#63](https://github.com/sveltejs/mcp/pull/63))
+- fix: minor tweaks to the prompt to allow for automatic sync ([#63](https://github.com/sveltejs/ai-tools/pull/63))
 
-- feat: `read_state_with_dollar` autofixer ([#66](https://github.com/sveltejs/mcp/pull/66))
+- feat: `read_state_with_dollar` autofixer ([#66](https://github.com/sveltejs/ai-tools/pull/66))
 
 ## 0.1.4
 
 ### Patch Changes
 
-- fix: pass secrets in action and update `mcpName` ([`044f098`](https://github.com/sveltejs/mcp/commit/044f0988b935fff39911a861a648dfb276f5831a))
+- fix: pass secrets in action and update `mcpName` ([`044f098`](https://github.com/sveltejs/ai-tools/commit/044f0988b935fff39911a861a648dfb276f5831a))
 
 ## 0.1.3
 
 ### Patch Changes
 
-- fix: use DNS to publish MCP ([#59](https://github.com/sveltejs/mcp/pull/59))
+- fix: use DNS to publish MCP ([#59](https://github.com/sveltejs/ai-tools/pull/59))
 
 ## 0.1.2
 
 ### Patch Changes
 
-- fix: publish to MCP registry (I really hope this time for real) ([`ef5241c`](https://github.com/sveltejs/mcp/commit/ef5241cbc204ad8bb84bde27db7c9d0a08280245))
+- fix: publish to MCP registry (I really hope this time for real) ([`ef5241c`](https://github.com/sveltejs/ai-tools/commit/ef5241cbc204ad8bb84bde27db7c9d0a08280245))
 
 ## 0.1.1
 
 ### Patch Changes
 
-- feat: publish mcp to registry (maybe for real this time) ([`132943d`](https://github.com/sveltejs/mcp/commit/132943db3b04dbbd322d08926c0880c990a61f5f))
+- feat: publish mcp to registry (maybe for real this time) ([`132943d`](https://github.com/sveltejs/ai-tools/commit/132943db3b04dbbd322d08926c0880c990a61f5f))
 
 ## 0.1.0
 
 ### Minor Changes
 
-- feat: publish to registry ([#45](https://github.com/sveltejs/mcp/pull/45))
+- feat: publish to registry ([#45](https://github.com/sveltejs/ai-tools/pull/45))
 
 ### Patch Changes
 
-- feat: add autofixer to tell the LLM to check if some function called in effect is assigning state #26 ([`73d7625`](https://github.com/sveltejs/mcp/commit/73d7625b3ca6a812ba91883ea668d80ff1e7c703))
+- feat: add autofixer to tell the LLM to check if some function called in effect is assigning state #26 ([`73d7625`](https://github.com/sveltejs/ai-tools/commit/73d7625b3ca6a812ba91883ea668d80ff1e7c703))
 
-- feat: add bind:this -> attachment and action -> attachment autofixer #20 ([`73d7625`](https://github.com/sveltejs/mcp/commit/73d7625b3ca6a812ba91883ea668d80ff1e7c703))
+- feat: add bind:this -> attachment and action -> attachment autofixer #20 ([`73d7625`](https://github.com/sveltejs/ai-tools/commit/73d7625b3ca6a812ba91883ea668d80ff1e7c703))
 
 ## 0.0.4
 
 ### Patch Changes
 
-- fix: allow TS `.svelte.ts` modules ([#49](https://github.com/sveltejs/mcp/pull/49))
+- fix: allow TS `.svelte.ts` modules ([#49](https://github.com/sveltejs/ai-tools/pull/49))
 
 ## 0.0.3
 
 ### Patch Changes
 
-- fix: check effect.pre in assign-in-effect ([#41](https://github.com/sveltejs/mcp/pull/41))
+- fix: check effect.pre in assign-in-effect ([#41](https://github.com/sveltejs/ai-tools/pull/41))
 
-- feat: `use_cases` documentation metadata ([#29](https://github.com/sveltejs/mcp/pull/29))
+- feat: `use_cases` documentation metadata ([#29](https://github.com/sveltejs/ai-tools/pull/29))
 
-- fix: change title names to allow for claude code to use the prompt ([`725f785`](https://github.com/sveltejs/mcp/commit/725f785766d04e9ed810a7c3f6bcfdb2e2b8234c))
+- fix: change title names to allow for claude code to use the prompt ([`725f785`](https://github.com/sveltejs/ai-tools/commit/725f785766d04e9ed810a7c3f6bcfdb2e2b8234c))
 
-- fix: enable doc tools ([`cb316c5`](https://github.com/sveltejs/mcp/commit/cb316c5b3ebc712946969d2d57236d159e796d58))
+- fix: enable doc tools ([`cb316c5`](https://github.com/sveltejs/ai-tools/commit/cb316c5b3ebc712946969d2d57236d159e796d58))
 
 ## 0.0.2
 
 ### Patch Changes
 
-- feat: latest version ([#25](https://github.com/sveltejs/mcp/pull/25))
+- feat: latest version ([#25](https://github.com/sveltejs/ai-tools/pull/25))

packages/mcp-stdio/checksums/registry_1.3.3_checksums.txt

@@ -1,25 +0,0 @@
-704f15b3ccba80990fbaaeb858e2e893120dea3ebb9340e04a36065987e089c5  mcp-publisher_darwin_amd64.tar.gz
-9eadadeb80998cd5b29e009dd3959634730d3754b9e201c0a3bb39dd6ae85933  mcp-publisher_darwin_amd64.tar.gz.sbom.json
-f80265eadc6b052d2885a7dfa47a1ecc7ca95926c143197ced747317793dcc11  mcp-publisher_darwin_arm64.tar.gz
-f063c462512eed8f2ce9610a6a125534cb552060ddb5364d8d49b32196416fbe  mcp-publisher_darwin_arm64.tar.gz.sbom.json
-1113b9d6bf59b000966c4f17752cf87b51db03dcc5482721421fd843ce3bf048  mcp-publisher_linux_amd64.tar.gz
-d3e46855b3b906aa84571b0d16384ad6074fd08709108a47d724908e91947b9d  mcp-publisher_linux_amd64.tar.gz.sbom.json
-34be13ec07490ab1250194d56820af222b49b85db6062ff848d33f9cf6eb41ce  mcp-publisher_linux_arm64.tar.gz
-a1ff3c26d53007c98301b747f97e5e241bfdb860b6a761db8627c9ba9ba71d88  mcp-publisher_linux_arm64.tar.gz.sbom.json
-9ba8b538744652c4837d08b7b83ae85d2dbea98c897cfcf4d90835df4fe075b3  mcp-publisher_windows_amd64.tar.gz
-924a0adcebd9360aed4ee5959e664e4e99b8259ed871ac91be5ace5657afba7d  mcp-publisher_windows_amd64.tar.gz.sbom.json
-a3eeab18ee6fd1d76d7ec153779ebe1bb404d7ea561f117781174e4396d90565  mcp-publisher_windows_arm64.tar.gz
-9351f3acc39f89402ce5f72be1b45f3bf1b16f8441c97496195188f276e6c227  mcp-publisher_windows_arm64.tar.gz.sbom.json
-9767b6f2afe1b60c597d4c66468d491be3466b25de8b4d4e7a21e61950856ef7  registry-1.3.3.tar.gz
-a7867ff2c15905a765dc77636d86241ea04a17d0a90f086913a3cea4fe899213  registry_darwin_amd64.tar.gz
-f0ed44b6ebdfcba40daefe7ca556f385de0cbae95dca3077c406f75cdb6abb9b  registry_darwin_amd64.tar.gz.sbom.json
-b4d93d6fc2d831d287cde79953822b434cb8406eea4f38f180628df227f6569e  registry_darwin_arm64.tar.gz
-dde0ed7f350eda98eb6f474d19aa0e4c717dd9f120609949acf05d7986cf6fcc  registry_darwin_arm64.tar.gz.sbom.json
-6772a573828f3482a192c88dac4223ee71ff4f521c8819480a194757e46a8272  registry_linux_amd64.tar.gz
-ac177354d5f43f86bdf7f36738837a2792ba136198209eb61d4d4dd9ab408853  registry_linux_amd64.tar.gz.sbom.json
-61c1fb9126977e0694d5aad8990a9d4c2d0e933e6aa469789559fad3865e8b7f  registry_linux_arm64.tar.gz
-9b81dfe2c81c49f7731072b8c217f05d609f7094ea94aaf9491f33880f2dec7f  registry_linux_arm64.tar.gz.sbom.json
-7fb5752de50a841614c6990f20807820bda33a72eaf7be51ae6cbd4e9acb2b87  registry_windows_amd64.tar.gz
-fd44139231d4725c75b2a5a96fb670d3137f628db91c8281f6348e842792d514  registry_windows_amd64.tar.gz.sbom.json
-a85ba7eb1a06a9e668b8a2097e00ab010e454a2afeb4e29ebeb7b702a31cccd3  registry_windows_arm64.tar.gz
-93503683d32a2971b741e87f2879cb2a3f41acd184c618e04d17a98926915f2d  registry_windows_arm64.tar.gz.sbom.json

packages/mcp-stdio/checksums/registry_1.4.0_checksums.txt

@@ -0,0 +1,25 @@
+eb5f89b76fc45a97070fa481eb03977584a78e3dff2781402d43482f114e4d6a  mcp-publisher_darwin_amd64.tar.gz
+29fc1b46a6f6be2580129369a9b681204b11b425d9a44147d79c8c658c7b8474  mcp-publisher_darwin_amd64.tar.gz.sbom.json
+9eddbbb95efd54b9503f6c0668f43bab3f04c856946d3c7164f6daead232402f  mcp-publisher_darwin_arm64.tar.gz
+a8349b0ea7916f34cf4ee4e1ced5b91bc1ded6d100312cbb2095da018710da04  mcp-publisher_darwin_arm64.tar.gz.sbom.json
+c4b402b43a85166c3f840641ca1c9e6de5bfa1cf533c22576d663ccbda0711bb  mcp-publisher_linux_amd64.tar.gz
+6f21cf917055be885f16b93154e06379540236599cfad6af4404a8323bde74b7  mcp-publisher_linux_amd64.tar.gz.sbom.json
+ba5d486f86b2cef48ea506e8314d901a5169dcd56a5d6e9daf18d41244316235  mcp-publisher_linux_arm64.tar.gz
+8a0096a407b916ac7270732df017a26f6112c73a066252f6556b8956c492a0b4  mcp-publisher_linux_arm64.tar.gz.sbom.json
+59ee8c4a997f94794db8db13f8809666631686a70a8d89a9f0fea993f9aede0f  mcp-publisher_windows_amd64.tar.gz
+57d211fd9181f698d126bd773b55c98b92454d19b1e32e77860766179a8a2e8e  mcp-publisher_windows_amd64.tar.gz.sbom.json
+1410952b0a5968cbe89590e7b4ee6105147ef7267cf0cd50095c9bec2ee3b0d7  mcp-publisher_windows_arm64.tar.gz
+6cb93e118a89ed1419135bfbaa7401bd3b7a7c5680a0d8fd7c78728f9d860630  mcp-publisher_windows_arm64.tar.gz.sbom.json
+ebc17c3b7a5b86f9c036acf1d44fb904bb363bad0ac1ac37b7979eb17cf3d218  registry-1.4.0.tar.gz
+5fffe8b078513fa5fbb625a213d164bb391c7c85e216e541cab789517bc6365b  registry_darwin_amd64.tar.gz
+10a61cf4173d8b5be63044af0a10e6c809eebc1006c0c1643753a252db808ddd  registry_darwin_amd64.tar.gz.sbom.json
+437746a1045f093266ad7298a47be41dc44cc33ecaeec145449c4eefeddf8880  registry_darwin_arm64.tar.gz
+4c0cb24bcef0658540fb044d771294efd662edecd2f7fae7b1ca7ca2ae68f83a  registry_darwin_arm64.tar.gz.sbom.json
+bd77fafcc881714a63a375a9bdb53a761a2b8e367a9d2759835126c993df2356  registry_linux_amd64.tar.gz
+d62a461b174089fbcaa4f1ac096bca491d18bc7f70e93ce0824fe89dbe42e974  registry_linux_amd64.tar.gz.sbom.json
+38cb9e6112ff11544ba8ec88c5c0f44d4c851504ec2e33d497e25616a6f7a21e  registry_linux_arm64.tar.gz
+84e5a004929e7231ae35350a1fe8fb08668fc05183e13d1d78004abf08a25f3b  registry_linux_arm64.tar.gz.sbom.json
+77ca9243d2f744f282b39d07625f802d77f593850a0392debabe407643a8579c  registry_windows_amd64.tar.gz
+bb043e8a6a8d187ffab8987d36d0018024115d9217af6d2ddd233f94aea880ea  registry_windows_amd64.tar.gz.sbom.json
+e4ee50e05a95f288b874f5e24c7716a5032548feeabc83b715193298cec06890  registry_windows_arm64.tar.gz
+49f189b615bce3d09ea24ae5d80e0816ac69225b3c8901dd7be19ef9ca06830c  registry_windows_arm64.tar.gz.sbom.json

packages/mcp-stdio/package.json

@@ -1,24 +1,30 @@
 {
 	"name": "@sveltejs/mcp",
-	"version": "0.1.10",
+	"version": "0.1.22",
 	"type": "module",
 	"license": "MIT",
 	"mcpName": "dev.svelte/mcp",
-	"homepage": "https://github.com/sveltejs/mcp#readme",
+	"homepage": "https://github.com/sveltejs/ai-tools#readme",
 	"bugs": {
-		"url": "https://github.com/sveltejs/mcp/issues"
+		"url": "https://github.com/sveltejs/ai-tools/issues"
 	},
 	"bin": {
-		"svelte-mcp": "./dist/index.js"
+		"svelte-mcp": "./dist/index.mjs"
 	},
 	"repository": {
 		"type": "git",
-		"url": "git+https://github.com/sveltejs/mcp.git",
+		"url": "git+https://github.com/sveltejs/ai-tools.git",
 		"path": "packages/mcp-stdio"
 	},
 	"files": [
 		"dist"
 	],
+	"exports": {
+		".": {
+			"types": "./dist/handlers.d.mts",
+			"import": "./dist/handlers.mjs"
+		}
+	},
 	"publishConfig": {
 		"access": "public"
 	},
@@ -32,15 +38,16 @@
 	},
 	"devDependencies": {
 		"@sveltejs/mcp-server": "workspace:^",
-		"@tmcp/transport-stdio": "^0.3.1",
-		"@types/node": "^22.15.17",
-		"publint": "^0.3.13",
-		"tsdown": "^0.15.0",
-		"typescript": "^5.8.3",
-		"vitest": "^4.0.0"
+		"@tmcp/transport-stdio": "catalog:tmcp",
+		"@types/node": "catalog:tooling",
+		"publint": "catalog:tooling",
+		"tsdown": "catalog:tooling",
+		"typescript": "catalog:tooling",
+		"vitest": "catalog:tooling"
 	},
 	"dependencies": {
-		"eslint": "^9.36.0",
-		"tmcp": "^1.15.3"
+		"eslint": "catalog:lint",
+		"sade": "catalog:tooling",
+		"tmcp": "catalog:tmcp"
 	}
 }

packages/mcp-stdio/server.json

@@ -1,14 +1,15 @@
 {
-	"$schema": "https://static.modelcontextprotocol.io/schemas/2025-10-17/server.schema.json",
+	"$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
 	"name": "dev.svelte/mcp",
+	"title": "Svelte MCP",
 	"description": "The official Svelte MCP server providing docs and autofixing tools for Svelte development",
 	"repository": {
 		"id": "1054419133",
-		"url": "https://github.com/sveltejs/mcp",
+		"url": "https://github.com/sveltejs/ai-tools",
 		"subfolder": "packages/mcp-stdio",
 		"source": "github"
 	},
-	"version": "0.1.10",
+	"version": "0.1.22",
 	"websiteUrl": "https://svelte.dev/docs/mcp/overview",
 	"icons": [
 		{
@@ -24,7 +25,7 @@
 		{
 			"registryType": "npm",
 			"identifier": "@sveltejs/mcp",
-			"version": "0.1.10",
+			"version": "0.1.22",
 			"runtimeHint": "npx",
 			"transport": {
 				"type": "stdio"

packages/mcp-stdio/src/handlers.ts

@@ -0,0 +1,6 @@
+export {
+	list_sections_handler as listSections,
+	get_documentation_handler as getDocumentation,
+	svelte_autofixer_handler as svelteAutofixer,
+	playground_link_handler as playgroundLink,
+} from '@sveltejs/mcp-server/handlers';

packages/mcp-stdio/src/index.ts

@@ -1,7 +1,65 @@
 #! /usr/bin/env node
 import { server } from '@sveltejs/mcp-server';
+import {
+	list_sections_handler,
+	get_documentation_handler,
+	svelte_autofixer_handler,
+} from '@sveltejs/mcp-server/handlers';
+import { readFile } from 'node:fs/promises';
+import { existsSync } from 'node:fs';
 import { StdioTransport } from '@tmcp/transport-stdio';
+import sade from 'sade';
 
-const transport = new StdioTransport(server);
+const cli = sade('svelte-mcp');
 
-transport.listen();
+cli.command('__mcp', '', { default: true }).action(() => {
+	const transport = new StdioTransport(server);
+	transport.listen();
+});
+
+cli
+	.command('list-sections')
+	.describe('List all the available documentation sections')
+	.action(async () => {
+		console.log(await list_sections_handler());
+	});
+
+cli
+	.command('get-documentation <sections>')
+	.describe('Get documentation for specified sections, separated by commas')
+	.action(async (sections) => {
+		console.log(await get_documentation_handler({ section: sections.split(',') }));
+	});
+
+cli
+	.command('svelte-autofixer <code_or_path>')
+	.describe(
+		'Detect and suggest fixes for Svelte code issues, because the terminal will substitute variables `$` should be correctly escaped',
+	)
+	.option('--async', 'Wether the project is using async svelte or not', false)
+	.option('--svelte-version', 'Which version of svelte to use...it can be 4 or 5', 5)
+	.action(async (code_or_path, { async, 'svelte-version': version }) => {
+		let code = code_or_path;
+
+		let is_path = false;
+
+		if (existsSync(code_or_path)) {
+			console.log('Detected file path, reading file...');
+			code = await readFile(code_or_path, 'utf-8');
+			is_path = true;
+		} else {
+			console.log('File not found, treating input as code...');
+		}
+
+		const desired_svelte_version = +version;
+
+		const result = await svelte_autofixer_handler({
+			code,
+			async: Boolean(async),
+			desired_svelte_version,
+			filename: is_path ? code_or_path : undefined,
+		});
+		console.log(result);
+	});
+
+cli.parse(process.argv);

packages/mcp-stdio/tsdown.config.ts

@@ -2,7 +2,7 @@ import { defineConfig } from 'tsdown';
 
 export default defineConfig([
 	{
-		entry: ['./src/index.ts'],
+		entry: ['./src/index.ts', './src/handlers.ts'],
 		platform: 'node',
 		define: {
 			// some eslint-plugin-svelte code expects __filename to exists but in an ESM environment it does not.
@@ -13,9 +13,12 @@ export default defineConfig([
 		// the require would fail once executed in a project without eslint installed.
 		external: ['eslint'],
 		publint: true,
-		dts: false,
+		dts: {
+			eager: true,
+		},
 		treeshake: true,
 		clean: true,
 		target: 'esnext',
+		inlineOnly: false,
 	},
 ]);

packages/opencode/CHANGELOG.md

@@ -0,0 +1,73 @@
+# @sveltejs/opencode
+
+## 0.1.8
+
+### Patch Changes
+
+- fix: add `server` export to opencode plugin ([`96c50ac`](https://github.com/sveltejs/ai-tools/commit/96c50acae2b4131a6c72d3579a73c44ab9158b18))
+
+## 0.1.7
+
+### Patch Changes
+
+- fix: import `ts` files directly ([#190](https://github.com/sveltejs/ai-tools/pull/190))
+
+## 0.1.6
+
+### Patch Changes
+
+- chore: sync skills from svelte.dev ([#178](https://github.com/sveltejs/ai-tools/pull/178))
+
+- fix: update svelte-file-editor agent to use proper name ([#183](https://github.com/sveltejs/ai-tools/pull/183))
+
+## 0.1.5
+
+### Patch Changes
+
+- fix: merge user-configured svelte-file-editor agent settings ([#176](https://github.com/sveltejs/ai-tools/pull/176))
+
+- feat(opencode): mcp enabled option is passed to opencode ([#171](https://github.com/sveltejs/ai-tools/pull/171))
+
+- feat: allow enabling a specific skill in opencode plugin ([#174](https://github.com/sveltejs/ai-tools/pull/174))
+
+## 0.1.4
+
+### Patch Changes
+
+- feat: `svelte-core-bestpractices` skill ([#162](https://github.com/sveltejs/ai-tools/pull/162))
+
+## 0.1.3
+
+### Patch Changes
+
+- fix: better subagent instructions to use MCP or skill ([#163](https://github.com/sveltejs/ai-tools/pull/163))
+
+## 0.1.2
+
+### Patch Changes
+
+- feat: allow for local opencode config ([#156](https://github.com/sveltejs/ai-tools/pull/156))
+
+## 0.1.1
+
+### Patch Changes
+
+- fix: actually push skills to right config path ([`c2c1b3e`](https://github.com/sveltejs/ai-tools/commit/c2c1b3e5e788b14eea17cd37a83ca55433cc4072))
+
+## 0.1.0
+
+### Minor Changes
+
+- feat: distribute skills through opencode plugin ([#151](https://github.com/sveltejs/ai-tools/pull/151))
+
+## 0.0.3
+
+### Patch Changes
+
+- chore: add README to opencode plugin ([`71295bc`](https://github.com/sveltejs/ai-tools/commit/71295bc11fb7bac6703e655f5fddead29967353c))
+
+## 0.0.2
+
+### Patch Changes
+
+- fix: show toast after a few seconds to avoid race condition ([`57e2d1d`](https://github.com/sveltejs/ai-tools/commit/57e2d1def1f5590d0a3dd6d269ac39f6397ffecf))

packages/opencode/README.md

@@ -0,0 +1,112 @@
+# @sveltejs/opencode
+
+OpenCode plugin for Svelte that provides the Svelte MCP server, a specialized file editor subagent and instruction files.
+
+## Installation
+
+Add `@sveltejs/opencode` to your OpenCode config (either global or local):
+
+```json
+{
+	"$schema": "https://opencode.ai/config.json",
+	"plugin": ["@sveltejs/opencode"]
+}
+```
+
+That's it! You now have the Svelte MCP server and the file editor subagent configured automatically.
+
+## Features
+
+### Svelte MCP Server
+
+The plugin automatically configures the [Svelte MCP server](https://mcp.svelte.dev) which provides:
+
+- **list-sections** - Discover available Svelte 5 and SvelteKit documentation sections
+- **get-documentation** - Retrieve full documentation content for specific sections
+- **svelte-autofixer** - Analyze Svelte code and get issues/suggestions
+- **playground-link** - Generate Svelte Playground links with provided code
+
+### Svelte File Editor Subagent
+
+A specialized subagent (`svelte-file-editor`) that is automatically used when creating, editing, or reviewing `.svelte`, `.svelte.ts`, or `.svelte.js` files. It fetches relevant documentation and validates code using the Svelte MCP server tools.
+
+### Agent Instructions
+
+The plugin injects instructions that teach the agent how to effectively use the Svelte MCP tools.
+
+## Configuration
+
+Create `svelte.json` to customize how the plugin configures MCP, the Svelte subagent, instructions, and skills.
+
+```json
+{
+	"$schema": "https://svelte.dev/opencode/schema.json",
+	"mcp": {
+		"type": "remote",
+		"enabled": true
+	},
+	"subagent": {
+		"enabled": true,
+		"agents": {
+			"svelte-file-editor": {
+				"model": "anthropic/claude-sonnet-4-20250514",
+				"temperature": 0.7,
+				"top_p": 0.9,
+				"maxSteps": 20
+			}
+		}
+	},
+	"instructions": {
+		"enabled": true
+	},
+	"skills": {
+		"enabled": ["svelte-code-writer", "svelte-core-bestpractices"]
+	}
+}
+```
+
+### Defaults
+
+If omitted, the plugin uses these defaults:
+
+- `mcp.type`: `"remote"`
+- `mcp.enabled`: `true`
+- `subagent.enabled`: `true`
+- `subagent.agents`: `{}`
+- `instructions.enabled`: `true`
+- `skills.enabled`: `true`
+
+### Configuration Options
+
+| Option                                           | Type                  | Default    | Description                                                                                    |
+| ------------------------------------------------ | --------------------- | ---------- | ---------------------------------------------------------------------------------------------- |
+| `mcp.type`                                       | `"remote" \| "local"` | `"remote"` | Use `https://mcp.svelte.dev/mcp` (`remote`) or run `@sveltejs/mcp` via `npx` (`local`).        |
+| `mcp.enabled`                                    | `boolean`             | `true`     | Enable or disable the Svelte MCP server entry.                                                 |
+| `subagent.enabled`                               | `boolean`             | `true`     | Enable or disable registration of the `svelte-file-editor` subagent.                           |
+| `subagent.agents.svelte-file-editor.model`       | `string`              | main agent | Override the model used by the Svelte file editor subagent.                                    |
+| `subagent.agents.svelte-file-editor.temperature` | `number`              | unset      | Set temperature for the subagent.                                                              |
+| `subagent.agents.svelte-file-editor.top_p`       | `number`              | unset      | Set top-p sampling for the subagent.                                                           |
+| `subagent.agents.svelte-file-editor.maxSteps`    | `number`              | unlimited  | Limit the number of steps the subagent can execute.                                            |
+| `instructions.enabled`                           | `boolean`             | `true`     | Enable or disable automatic instruction-file injection.                                        |
+| `skills.enabled`                                 | `boolean \| string[]` | `true`     | Enable all skills (`true`), disable all skills (`false`), or enable only specific skill names. |
+
+### Supported Skill Names
+
+When using `skills.enabled` as an array, these built-in names are currently available:
+
+- `svelte-code-writer`
+- `svelte-core-bestpractices`
+
+### Config File Locations and Precedence
+
+The plugin reads from these files (lowest priority first, highest priority last):
+
+- `~/.config/opencode/svelte.json`
+- `$OPENCODE_CONFIG_DIR/svelte.json` (when `OPENCODE_CONFIG_DIR` is set)
+- `.opencode/svelte.json` in the current project
+
+If the same key is defined in multiple files, the later location overrides earlier ones.
+
+## License
+
+MIT

packages/opencode/agents.ts

@@ -0,0 +1,11 @@
+// This file is auto-generated by scripts/sync-opencode-plugin.ts
+// Do not edit manually — edit the markdown files in tools/agents/ instead.
+
+export const agents = {
+	'svelte-file-editor': {
+		description:
+			'Specialized Svelte 5 code editor. MUST BE USED PROACTIVELY when creating, editing, or reviewing any .svelte file or .svelte.ts/.svelte.js module and MUST use the tools from the MCP server or the `svelte-file-editor` skill if they are available. Fetches relevant documentation and validates code using the Svelte MCP server tools.',
+		prompt:
+			"You are a Svelte 5 expert responsible for writing, editing, and validating Svelte components and modules. You have access to the Svelte MCP server which provides documentation and code analysis tools. Always use the tools from the svelte MCP server to fetch documentation with `get_documentation` and validating the code with `svelte_autofixer`. If the autofixer returns any issue or suggestions try to solve them.\n\nIf the MCP tools are not available you can use the `svelte-code-writer` skill to learn how to use the `@sveltejs/mcp` cli to access the same tools.\n\nIf the skill is not available you can run `npx @sveltejs/mcp@latest -y --help` to learn how to use it.\n\n## Available MCP Tools\n\n### 1. list-sections\n\nLists all available Svelte 5 and SvelteKit documentation sections with titles and paths. Use this first to discover what documentation is available.\n\n### 2. get-documentation\n\nRetrieves full documentation for specified sections. Accepts a single section name or an array of section names. Use after `list-sections` to fetch relevant docs for the task at hand.\n\n**Example sections:** `$state`, `$derived`, `$effect`, `$props`, `$bindable`, `snippets`, `routing`, `load functions`\n\n### 3. svelte-autofixer\n\nAnalyzes Svelte code and returns suggestions to fix issues. Pass the component code directly to this tool. It will detect common mistakes like:\n\n- Using `$effect` instead of `$derived` for computations\n- Missing cleanup in effects\n- Svelte 4 syntax (`on:click`, `export let`, `<slot>`)\n- Missing keys in `{#each}` blocks\n- And more\n\n## Workflow\n\nWhen invoked to work on a Svelte file:\n\n### 1. Gather Context (if needed)\n\nIf you're uncertain about Svelte 5 syntax or patterns, use the MCP tools:\n\n1. Call `list-sections` to see available documentation\n2. Call `get-documentation` with relevant section names\n\n### 2. Read the Target File\n\nRead the file to understand the current implementation.\n\n### 3. Make Changes\n\nApply edits following Svelte 5 best practices:\n\n### 4. Validate Changes\n\nAfter editing, ALWAYS call `svelte-autofixer` with the updated code to check for issues.\n\n### 5. Fix Any Issues\n\nIf the autofixer reports problems, fix them and re-validate until no issues remain.\n\n## Output Format\n\nAfter completing your work, provide:\n\n1. Summary of changes made\n2. Any issues found and fixed by the autofixer\n3. Recommendations for further improvements (if any)",
+	},
+} as const;

packages/opencode/config.ts

@@ -0,0 +1,229 @@
+import type { PluginInput } from '@opencode-ai/plugin';
+import { existsSync, readFileSync } from 'fs';
+import { homedir } from 'os';
+import { join } from 'path';
+import * as v from 'valibot';
+
+// Schema for individual agent configuration
+const agent_config_schema = v.object({
+	model: v.pipe(
+		v.optional(v.string()),
+		v.description('Model identifier for the agent (e.g., "anthropic/claude-sonnet-4-20250514")'),
+	),
+	temperature: v.pipe(
+		v.optional(v.number()),
+		v.description('Temperature setting for the agent (e.g., 0.7)'),
+	),
+	top_p: v.pipe(
+		v.optional(v.number()),
+		v.description(
+			'Control response diversity with the top_p option. Alternative to temperature for controlling randomness.',
+		),
+	),
+	maxSteps: v.pipe(
+		v.optional(v.number()),
+		v.description('Maximum number of steps the agent can take (e.g., 10)'),
+	),
+});
+
+const default_config = {
+	mcp: {
+		type: 'remote' as 'remote' | 'local',
+		enabled: true,
+	},
+	subagent: {
+		enabled: true,
+		agents: {} as Record<string, v.InferInput<typeof agent_config_schema>>,
+	},
+	instructions: {
+		enabled: true,
+	},
+	skills: {
+		enabled: true as boolean | string[],
+	},
+};
+
+export const config_schema = v.object({
+	mcp: v.pipe(
+		v.optional(
+			v.object({
+				type: v.optional(v.picklist(['remote', 'local'])),
+				enabled: v.optional(v.boolean()),
+			}),
+		),
+		v.description(
+			"Configuration for the MCP. You can chose if it should be enabled or not and the transport to use 'remote' (default) and 'local'.",
+		),
+	),
+	subagent: v.pipe(
+		v.optional(
+			v.object({
+				enabled: v.optional(v.boolean()),
+				agents: v.optional(v.record(v.string(), agent_config_schema)),
+			}),
+		),
+		v.description('Configuration for the subagent. You can choose if it should be enabled or not.'),
+	),
+	instructions: v.pipe(
+		v.optional(
+			v.object({
+				enabled: v.optional(v.boolean()),
+			}),
+		),
+		v.description(
+			'Configuration for the automatic AGENTS.md injection. You can choose if it should be enabled or not.',
+		),
+	),
+	skills: v.pipe(
+		v.optional(
+			v.object({
+				enabled: v.pipe(
+					v.optional(v.union([v.boolean(), v.array(v.string())])),
+					v.description(
+						'It can be either a boolean or an array containing the skills that you want to enable',
+					),
+				),
+			}),
+		),
+		v.description(
+			'Configuration for the skills. You can choose if it they should be enabled or not, or specify an array of skill names to enable only specific skills.',
+		),
+	),
+});
+
+export type McpConfig = v.InferInput<typeof config_schema>;
+
+const GLOBAL_CONFIG_DIR = join(homedir(), '.config', 'opencode');
+const GLOBAL_CONFIG_PATH = join(GLOBAL_CONFIG_DIR, 'svelte.json');
+
+interface ConfigLoadResult {
+	data: Record<string, unknown> | null;
+	parse_error?: string;
+}
+
+function get_config_paths() {
+	// Global: ~/.config/opencode/svelte.json
+	let global_path: string | null = null;
+	if (existsSync(GLOBAL_CONFIG_PATH)) {
+		global_path = GLOBAL_CONFIG_PATH;
+	}
+
+	// Custom config directory: $OPENCODE_CONFIG_DIR/svelte.json
+	let config_dir_path: string | null = null;
+	const opencode_config_dir = process.env.OPENCODE_CONFIG_DIR;
+	if (opencode_config_dir) {
+		const config_json = join(opencode_config_dir, 'svelte.json');
+		if (existsSync(config_json)) {
+			config_dir_path = config_json;
+		}
+	}
+
+	// Project-local: ./.opencode/svelte.json (cwd)
+	let project_path: string | null = null;
+	const project_config = join(process.cwd(), '.opencode', 'svelte.json');
+	if (existsSync(project_config)) {
+		project_path = project_config;
+	}
+
+	// Lowest priority first, highest priority last (project overrides global)
+	return [global_path, config_dir_path, project_path];
+}
+
+function load_config_file(config_path: string): ConfigLoadResult {
+	let file_content: string;
+	try {
+		file_content = readFileSync(config_path, 'utf-8');
+	} catch {
+		// File doesn't exist or can't be read
+		return { data: null };
+	}
+
+	try {
+		const parsed = JSON.parse(file_content);
+		if (parsed === undefined || parsed === null) {
+			return { data: null, parse_error: 'Config file is empty or invalid' };
+		}
+		return { data: parsed };
+	} catch (error: unknown) {
+		return {
+			data: null,
+			parse_error: error instanceof Error ? error.message : 'Failed to parse config',
+		};
+	}
+}
+
+function merge_with_defaults(user_config: Partial<McpConfig>): McpConfig {
+	return {
+		mcp: {
+			...default_config.mcp,
+			...user_config.mcp,
+		},
+		subagent: {
+			enabled: default_config.subagent.enabled,
+			...user_config.subagent,
+			agents: {
+				...default_config.subagent.agents,
+				...user_config.subagent?.agents,
+			},
+		},
+		instructions: {
+			...default_config.instructions,
+			...user_config.instructions,
+		},
+		skills: {
+			...default_config.skills,
+			...user_config.skills,
+		},
+	};
+}
+
+export function get_mcp_config(ctx: PluginInput) {
+	const config_paths = get_config_paths();
+	let merged: Partial<McpConfig> = {};
+
+	// Iterate from lowest to highest priority, merging as we go
+	for (const path of config_paths) {
+		if (path && existsSync(path)) {
+			const result = load_config_file(path);
+			if (result.parse_error) {
+				setTimeout(() => {
+					ctx.client.tui.showToast({
+						body: {
+							title: 'Svelte: Invalid opencode plugin config',
+							message: `${result.parse_error} (${path})\nSkipping this config file`,
+							variant: 'warning',
+							duration: 7000,
+						},
+					});
+				}, 7000);
+				continue;
+			}
+			const parsed = v.safeParse(config_schema, result.data);
+			if (parsed.success) {
+				merged = {
+					mcp: { ...merged.mcp, ...parsed.output.mcp },
+					subagent: {
+						...merged.subagent,
+						...parsed.output.subagent,
+						agents: { ...merged.subagent?.agents, ...parsed.output.subagent?.agents },
+					},
+					instructions: { ...merged.instructions, ...parsed.output.instructions },
+					skills: { ...merged.skills, ...parsed.output.skills },
+				};
+			} else {
+				setTimeout(() => {
+					ctx.client.tui.showToast({
+						body: {
+							title: 'Svelte: Invalid opencode plugin config',
+							message: `Invalid config schema (${path})\nSkipping this config file`,
+							variant: 'warning',
+							duration: 7000,
+						},
+					});
+				}, 7000);
+			}
+		}
+	}
+
+	return merge_with_defaults(merged);
+}

packages/opencode/index.ts

@@ -0,0 +1,117 @@
+import type { Plugin } from '@opencode-ai/plugin';
+import { readdir } from 'node:fs/promises';
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { agents } from './agents.ts';
+import { get_mcp_config } from './config.ts';
+
+const current_dir = dirname(fileURLToPath(import.meta.url));
+
+export const svelte_plugin: Plugin = async (ctx) => {
+	return {
+		async config(input) {
+			input.agent ??= {};
+			input.mcp ??= {};
+			input.instructions ??= [];
+			// @ts-expect-error -- types are wrong in the opencode package...will fix there and remove this
+			input.skills ??= {};
+			// @ts-expect-error -- types are wrong in the opencode package...will fix there and remove this
+			input.skills.paths ??= [];
+			// by default we use svelte as the name for the svelte MCP server
+			let svelte_mcp_name = 'svelte';
+			// we loop over every mcp server to see if any of them is already the svelte MCP server
+			for (const name in input.mcp) {
+				const mcp = input.mcp[name];
+				if (
+					(mcp?.type === 'remote' && mcp.url.includes('https://mcp.svelte.dev/mcp')) ||
+					(mcp?.type === 'local' &&
+						mcp.command.some((cmd: string) => cmd.includes('@sveltejs/mcp')))
+				) {
+					// if we found the svelte MCP server, we store its name and break
+					svelte_mcp_name = name;
+					break;
+				}
+			}
+			const mcp_config = get_mcp_config(ctx);
+
+			if (mcp_config.instructions?.enabled !== false) {
+				const instructions_dir = join(current_dir, 'instructions');
+				const instructions_paths = await readdir(instructions_dir);
+				input.instructions.push(...instructions_paths.map((file) => join(instructions_dir, file)));
+			}
+
+			const skills_enabled = mcp_config.skills?.enabled;
+			if (skills_enabled !== false) {
+				const skills_dir = join(current_dir, 'skills');
+				if (Array.isArray(skills_enabled)) {
+					// only add specific skill directories by name
+					for (const skill_name of skills_enabled) {
+						const skill_path = join(skills_dir, skill_name);
+						// @ts-expect-error -- skills is a new opencode feature
+						input.skills.paths.push(skill_path);
+					}
+				} else {
+					// @ts-expect-error -- skills is a new opencode feature
+					input.skills.paths.push(skills_dir);
+				}
+			}
+
+			// if the user doesn't have the MCP server already we add one based on config
+			if (!input.mcp[svelte_mcp_name]) {
+				if (mcp_config.mcp?.type === 'remote') {
+					input.mcp[svelte_mcp_name] = {
+						type: 'remote',
+						url: 'https://mcp.svelte.dev/mcp',
+						enabled: mcp_config.mcp?.enabled ?? true,
+					};
+				} else {
+					input.mcp[svelte_mcp_name] = {
+						type: 'local',
+						command: ['npx', '-y', '@sveltejs/mcp'],
+						enabled: mcp_config.mcp?.enabled ?? true,
+					};
+				}
+			}
+			if (mcp_config.subagent?.enabled !== false) {
+				for (const [agent_name, agent_data] of Object.entries(agents)) {
+					// we add the editor subagent that will be used when editing Svelte files to prevent wasting context on the main agent
+					const default_config: (typeof input.agent)[string] = {
+						color: '#ff3e00',
+						mode: 'subagent',
+						prompt: agent_data.prompt,
+						description: agent_data.description,
+						permission: {
+							bash: 'ask',
+							edit: 'allow',
+							webfetch: 'ask',
+						},
+						tools: {
+							[`${svelte_mcp_name}_*`]: true,
+						},
+					};
+
+					// Get per-agent config from svelte.json (if any)
+					const agent_config = mcp_config.subagent?.agents?.[agent_name];
+
+					// Configure agent from svelte.json only
+					// Priority: svelte.json agent config > defaults
+					input.agent[agent_name] = {
+						...default_config,
+						...(agent_config?.model !== undefined && {
+							model: agent_config.model,
+						}),
+						...(agent_config?.temperature !== undefined && {
+							temperature: agent_config.temperature,
+						}),
+						...(agent_config?.maxSteps !== undefined && {
+							maxSteps: agent_config.maxSteps,
+						}),
+						...(agent_config?.top_p !== undefined && {
+							top_p: agent_config.top_p,
+						}),
+					};
+				}
+			}
+		},
+	};
+};

packages/opencode/instructions/opencode-agents.md

@@ -0,0 +1,23 @@
+You are able to use the Svelte MCP server, where you have access to comprehensive Svelte 5 and SvelteKit documentation. Here's how to use the available tools effectively:
+
+## Available Svelte MCP Tools:
+
+### 1. list-sections
+
+Use this FIRST to discover all available documentation sections. Returns a structured list with titles, use_cases, and paths.
+When asked about Svelte or SvelteKit topics, ALWAYS use this tool at the start of the chat to find relevant sections.
+
+### 2. get-documentation
+
+Retrieves full documentation content for specific sections. Accepts single or multiple sections.
+After calling the list-sections tool, you MUST analyze the returned documentation sections (especially the use_cases field) and then use the get-documentation tool to fetch ALL documentation sections that are relevant for the user's task.
+
+### 3. svelte-autofixer
+
+Analyzes Svelte code and returns issues and suggestions.
+You MUST use this tool whenever writing Svelte code before sending it to the user. Keep calling it until no issues or suggestions are returned.
+
+### 4. playground-link
+
+Generates a Svelte Playground link with the provided code.
+After completing the code, ask the user if they want a playground link. Only call this tool after user confirmation and NEVER if code was written to files in their project.

packages/opencode/package.json

@@ -0,0 +1,43 @@
+{
+	"name": "@sveltejs/opencode",
+	"version": "0.1.8",
+	"type": "module",
+	"license": "MIT",
+	"homepage": "https://github.com/sveltejs/ai-tools#readme",
+	"bugs": {
+		"url": "https://github.com/sveltejs/ai-tools/issues"
+	},
+	"scripts": {
+		"check": "tsc --noEmit",
+		"generate-schema": "node --import node-resolve-ts/register scripts/generate-schema.ts"
+	},
+	"files": [
+		"index.ts",
+		"config.ts",
+		"agents.ts",
+		"instructions",
+		"skills"
+	],
+	"exports": {
+		"./server": {
+			"types": "./index.ts",
+			"import": "./index.ts"
+		}
+	},
+	"repository": {
+		"type": "git",
+		"url": "git+https://github.com/sveltejs/ai-tools.git",
+		"path": "packages/opencode"
+	},
+	"publishConfig": {
+		"access": "public"
+	},
+	"dependencies": {
+		"valibot": "catalog:tooling"
+	},
+	"devDependencies": {
+		"@opencode-ai/plugin": "catalog:ai",
+		"@valibot/to-json-schema": "catalog:tooling",
+		"@types/node": "catalog:tooling"
+	}
+}

packages/opencode/schema.json

@@ -0,0 +1,111 @@
+{
+	"type": "object",
+	"properties": {
+		"mcp": {
+			"type": "object",
+			"properties": {
+				"type": {
+					"enum": [
+						"remote",
+						"local"
+					]
+				},
+				"enabled": {
+					"type": "boolean"
+				}
+			},
+			"required": [],
+			"description": "Configuration for the MCP. You can chose if it should be enabled or not and the transport to use 'remote' (default) and 'local'."
+		},
+		"subagent": {
+			"type": "object",
+			"properties": {
+				"enabled": {
+					"type": "boolean"
+				},
+				"agents": {
+					"type": "object",
+					"propertyNames": {
+						"anyOf": [
+							{
+								"enum": [
+									"svelte-file-editor"
+								]
+							},
+							{
+								"type": "string"
+							}
+						]
+					},
+					"additionalProperties": {
+						"type": "object",
+						"properties": {
+							"model": {
+								"type": "string",
+								"description": "Model identifier for the agent (e.g., \"anthropic/claude-sonnet-4-20250514\")"
+							},
+							"temperature": {
+								"type": "number",
+								"description": "Temperature setting for the agent (e.g., 0.7)"
+							},
+							"top_p": {
+								"type": "number",
+								"description": "Control response diversity with the top_p option. Alternative to temperature for controlling randomness."
+							},
+							"maxSteps": {
+								"type": "number",
+								"description": "Maximum number of steps the agent can take (e.g., 10)"
+							}
+						},
+						"required": []
+					}
+				}
+			},
+			"required": [],
+			"description": "Configuration for the subagent. You can choose if it should be enabled or not."
+		},
+		"instructions": {
+			"type": "object",
+			"properties": {
+				"enabled": {
+					"type": "boolean"
+				}
+			},
+			"required": [],
+			"description": "Configuration for the automatic AGENTS.md injection. You can choose if it should be enabled or not."
+		},
+		"skills": {
+			"type": "object",
+			"properties": {
+				"enabled": {
+					"anyOf": [
+						{
+							"type": "boolean"
+						},
+						{
+							"type": "array",
+							"items": {
+								"anyOf": [
+									{
+										"enum": [
+											"svelte-code-writer",
+											"svelte-core-bestpractices"
+										]
+									},
+									{
+										"type": "string"
+									}
+								]
+							}
+						}
+					],
+					"description": "It can be either a boolean or an array containing the skills that you want to enable"
+				}
+			},
+			"required": [],
+			"description": "Configuration for the skills. You can choose if it they should be enabled or not, or specify an array of skill names to enable only specific skills."
+		}
+	},
+	"required": [],
+	"$schema": "http://json-schema.org/draft-07/schema#"
+}

packages/opencode/scripts/generate-schema.ts

@@ -0,0 +1,60 @@
+import { toJsonSchema } from '@valibot/to-json-schema';
+import { config_schema } from '../config.js';
+import fs from 'node:fs';
+import path from 'node:path';
+
+// Read agent names from tools/agents/*.md files
+function get_agent_names(agents_dir: string) {
+	if (!fs.existsSync(agents_dir)) return [];
+	return fs
+		.readdirSync(agents_dir, { withFileTypes: true })
+		.filter((entry) => entry.isFile() && entry.name.endsWith('.md'))
+		.map((entry) => entry.name.replace(/\.md$/, ''));
+}
+
+function get_skill_names(skills_dir: string) {
+	if (!fs.existsSync(skills_dir)) return [];
+	return fs
+		.readdirSync(skills_dir, { withFileTypes: true })
+		.filter((entry) => entry.isDirectory())
+		.map((entry) => entry.name);
+}
+
+const skills_dir = path.resolve('./skills');
+const skill_names = get_skill_names(skills_dir);
+const schema = config_schema;
+const json_schema = toJsonSchema(schema);
+
+// Post-process: inject skill name suggestions into the items schema.
+// This is the JSON Schema equivalent of `"a" | "b" | (string & {})` —
+// editors will autocomplete the known names but any string is still valid.
+if (skill_names.length > 0) {
+	// eslint-disable-next-line @typescript-eslint/no-explicit-any
+	const enabled = (json_schema as any).properties?.skills?.properties?.enabled;
+	if (enabled?.anyOf) {
+		const array_branch = enabled.anyOf.find((s: Record<string, unknown>) => s.type === 'array');
+		if (array_branch) {
+			array_branch.items = {
+				anyOf: [{ enum: skill_names }, { type: 'string' }],
+			};
+		}
+	}
+}
+
+// Post-process: inject known agent names for intellisense
+// This is the JSON Schema equivalent of `"a" | "b" | (string & {})` —
+// editors will autocomplete the known names but any string is still valid.
+const agents_dir = path.resolve('../../tools/agents');
+const agent_names = get_agent_names(agents_dir);
+
+if (agent_names.length > 0) {
+	// eslint-disable-next-line @typescript-eslint/no-explicit-any
+	const agents = (json_schema as any).properties?.subagent?.properties?.agents;
+	if (agents) {
+		agents.propertyNames = {
+			anyOf: [{ enum: agent_names }, { type: 'string' }],
+		};
+	}
+}
+
+fs.writeFileSync(path.resolve('./schema.json'), JSON.stringify(json_schema, null, '\t'));

packages/opencode/skills/svelte-code-writer/SKILL.md

@@ -0,0 +1,66 @@
+---
+name: svelte-code-writer
+description: CLI tools for Svelte 5 documentation lookup and code analysis. MUST be used whenever creating, editing or analyzing any Svelte component (.svelte) or Svelte module (.svelte.ts/.svelte.js). If possible, this skill should be executed within the svelte-file-editor agent for optimal results.
+---
+
+# Svelte 5 Code Writer
+
+## CLI Tools
+
+You have access to `@sveltejs/mcp` CLI for Svelte-specific assistance. Use these commands via `npx`:
+
+### List Documentation Sections
+
+```bash
+npx @sveltejs/mcp list-sections
+```
+
+Lists all available Svelte 5 and SvelteKit documentation sections with titles and paths.
+
+### Get Documentation
+
+```bash
+npx @sveltejs/mcp get-documentation "<section1>,<section2>,..."
+```
+
+Retrieves full documentation for specified sections. Use after `list-sections` to fetch relevant docs.
+
+**Example:**
+
+```bash
+npx @sveltejs/mcp get-documentation "$state,$derived,$effect"
+```
+
+### Svelte Autofixer
+
+```bash
+npx @sveltejs/mcp svelte-autofixer "<code_or_path>" [options]
+```
+
+Analyzes Svelte code and suggests fixes for common issues.
+
+**Options:**
+
+- `--async` - Enable async Svelte mode (default: false)
+- `--svelte-version` - Target version: 4 or 5 (default: 5)
+
+**Examples:**
+
+```bash
+# Analyze inline code (escape $ as \$)
+npx @sveltejs/mcp svelte-autofixer '<script>let count = \$state(0);</script>'
+
+# Analyze a file
+npx @sveltejs/mcp svelte-autofixer ./src/lib/Component.svelte
+
+# Target Svelte 4
+npx @sveltejs/mcp svelte-autofixer ./Component.svelte --svelte-version 4
+```
+
+**Important:** When passing code with runes (`$state`, `$derived`, etc.) via the terminal, escape the `$` character as `\$` to prevent shell variable substitution.
+
+## Workflow
+
+1. **Uncertain about syntax?** Run `list-sections` then `get-documentation` for relevant topics
+2. **Reviewing/debugging?** Run `svelte-autofixer` on the code to detect issues
+3. **Always validate** - Run `svelte-autofixer` before finalizing any Svelte component

packages/opencode/skills/svelte-core-bestpractices/SKILL.md

@@ -0,0 +1,176 @@
+---
+name: svelte-core-bestpractices
+description: Guidance on writing fast, robust, modern Svelte code. Load this skill whenever in a Svelte project and asked to write/edit or analyze a Svelte component or module. Covers reactivity, event handling, styling, integration with libraries and more.
+---
+
+## `$state`
+
+Only use the `$state` rune for variables that should be _reactive_ — in other words, variables that cause an `$effect`, `$derived` or template expression to update. Everything else can be a normal variable.
+
+Objects and arrays (`$state({...})` or `$state([...])`) are made deeply reactive, meaning mutation will trigger updates. This has a trade-off: in exchange for fine-grained reactivity, the objects must be proxied, which has performance overhead. In cases where you're dealing with large objects that are only ever reassigned (rather than mutated), use `$state.raw` instead. This is often the case with API responses, for example.
+
+## `$derived`
+
+To compute something from state, use `$derived` rather than `$effect`:
+
+```js
+// do this
+let square = $derived(num * num);
+
+// don't do this
+let square;
+
+$effect(() => {
+	square = num * num;
+});
+```
+
+> [!NOTE] `$derived` is given an expression, _not_ a function. If you need to use a function (because the expression is complex, for example) use `$derived.by`.
+
+Deriveds are writable — you can assign to them, just like `$state`, except that they will re-evaluate when their expression changes.
+
+If the derived expression is an object or array, it will be returned as-is — it is _not_ made deeply reactive. You can, however, use `$state` inside `$derived.by` in the rare cases that you need this.
+
+## `$effect`
+
+Effects are an escape hatch and should mostly be avoided. In particular, avoid updating state inside effects.
+
+- If you need to sync state to an external library such as D3, it is often neater to use [`{@attach ...}`](references/@attach.md)
+- If you need to run some code in response to user interaction, put the code directly in an event handler or use a [function binding](references/bind.md) as appropriate
+- If you need to log values for debugging purposes, use [`$inspect`](references/$inspect.md)
+- If you need to observe something external to Svelte, use [`createSubscriber`](references/svelte-reactivity.md)
+
+Never wrap the contents of an effect in `if (browser) {...}` or similar — effects do not run on the server.
+
+## `$props`
+
+Treat props as though they will change. For example, values that depend on props should usually use `$derived`:
+
+```js
+// @errors: 2451
+let { type } = $props();
+
+// do this
+let color = $derived(type === 'danger' ? 'red' : 'green');
+
+// don't do this — `color` will not update if `type` changes
+let color = type === 'danger' ? 'red' : 'green';
+```
+
+## `$inspect.trace`
+
+`$inspect.trace` is a debugging tool for reactivity. If something is not updating properly or running more than it should you can add `$inspect.trace(label)` as the first line of an `$effect` or `$derived.by` (or any function they call) to trace their dependencies and discover which one triggered an update.
+
+## Events
+
+Any element attribute starting with `on` is treated as an event listener:
+
+```svelte
+<button onclick={() => {...}}>click me</button>
+
+<!-- attribute shorthand also works -->
+<button {onclick}>...</button>
+
+<!-- so do spread attributes -->
+<button {...props}>...</button>
+```
+
+If you need to attach listeners to `window` or `document` you can use `<svelte:window>` and `<svelte:document>`:
+
+```svelte
+<svelte:window onkeydown={...} />
+<svelte:document onvisibilitychange={...} />
+```
+
+Avoid using `onMount` or `$effect` for this.
+
+## Snippets
+
+[Snippets](references/snippet.md) are a way to define reusable chunks of markup that can be instantiated with the [`{@render ...}`](references/@render.md) tag, or passed to components as props. They must be declared within the template.
+
+```svelte
+{#snippet greeting(name)}
+	<p>hello {name}!</p>
+{/snippet}
+
+{@render greeting('world')}
+```
+
+> [!NOTE] Snippets declared at the top level of a component (i.e. not inside elements or blocks) can be referenced inside `<script>`. A snippet that doesn't reference component state is also available in a `<script module>`, in which case it can be exported for use by other components.
+
+## Each blocks
+
+Prefer to use [keyed each blocks](references/each.md) — this improves performance by allowing Svelte to surgically insert or remove items rather than updating the DOM belonging to existing items.
+
+> [!NOTE] The key _must_ uniquely identify the object. Do not use the index as a key.
+
+Avoid destructuring if you need to mutate the item (with something like `bind:value={item.count}`, for example).
+
+## Using JavaScript variables in CSS
+
+If you have a JS variable that you want to use inside CSS you can set a CSS custom property with the `style:` directive.
+
+```svelte
+<div style:--columns={columns}>...</div>
+```
+
+You can then reference `var(--columns)` inside the component's `<style>`.
+
+## Styling child components
+
+The CSS in a component's `<style>` is scoped to that component. If a parent component needs to control the child's styles, the preferred way is to use CSS custom properties:
+
+```svelte
+<!-- Parent.svelte -->
+<Child --color="red" />
+
+<!-- Child.svelte -->
+<h1>Hello</h1>
+
+<style>
+	h1 {
+		color: var(--color);
+	}
+</style>
+```
+
+If this is impossible (for example, the child component comes from a library) you can use `:global` to override styles:
+
+```svelte
+<div>
+	<Child />
+</div>
+
+<style>
+	div :global {
+		h1 {
+			color: red;
+		}
+	}
+</style>
+```
+
+## Context
+
+Consider using context instead of declaring state in a shared module. This will scope the state to the part of the app that needs it, and eliminate the possibility of it leaking between users when server-side rendering.
+
+Use `createContext` rather than `setContext` and `getContext`, as it provides type safety.
+
+## Async Svelte
+
+If using version 5.36 or higher, you can use [await expressions](references/await-expressions.md) and [hydratable](references/hydratable.md) to use promises directly inside components. Note that these require the `experimental.async` option to be enabled in `svelte.config.js` as they are not yet considered fully stable.
+
+## Avoid legacy features
+
+Always use runes mode for new code, and avoid features that have more modern replacements:
+
+- use `$state` instead of implicit reactivity (e.g. `let count = 0; count += 1`)
+- use `$derived` and `$effect` instead of `$:` assignments and statements (but only use effects when there is no better solution)
+- use `$props` instead of `export let`, `$$props` and `$$restProps`
+- use `onclick={...}` instead of `on:click={...}`
+- use `{#snippet ...}` and `{@render ...}` instead of `<slot>` and `$$slots` and `<svelte:fragment>`
+- use `<DynamicComponent>` instead of `<svelte:component this={DynamicComponent}>`
+- use `import Self from './ThisComponent.svelte'` and `<Self>` instead of `<svelte:self>`
+- use classes with `$state` fields to share reactivity between components, instead of using stores
+- use `{@attach ...}` instead of `use:action`
+- use clsx-style arrays and objects in `class` attributes, instead of the `class:` directive

packages/opencode/skills/svelte-core-bestpractices/references/$inspect.md

@@ -0,0 +1,53 @@
+> [!NOTE] `$inspect` only works during development. In a production build it becomes a noop.
+
+The `$inspect` rune is roughly equivalent to `console.log`, with the exception that it will re-run whenever its argument changes. `$inspect` tracks reactive state deeply, meaning that updating something inside an object or array using fine-grained reactivity will cause it to re-fire (demo:
+
+```svelte
+<script>
+	let count = $state(0);
+	let message = $state('hello');
+
+	$inspect(count, message); // will console.log when `count` or `message` change
+</script>
+
+<button onclick={() => count++}>Increment</button>
+<input bind:value={message} />
+```
+
+On updates, a stack trace will be printed, making it easy to find the origin of a state change (unless you're in the playground, due to technical limitations).
+
+## $inspect(...).with
+
+`$inspect` returns a property `with`, which you can invoke with a callback, which will then be invoked instead of `console.log`. The first argument to the callback is either `"init"` or `"update"`; subsequent arguments are the values passed to `$inspect` (demo:
+
+```svelte
+<script>
+	let count = $state(0);
+
+	$inspect(count).with((type, count) => {
+		if (type === 'update') {
+			debugger; // or `console.trace`, or whatever you want
+		}
+	});
+</script>
+
+<button onclick={() => count++}>Increment</button>
+```
+
+## $inspect.trace(...)
+
+This rune, added in 5.14, causes the surrounding function to be _traced_ in development. Any time the function re-runs as part of an [effect]($effect) or a [derived]($derived), information will be printed to the console about which pieces of reactive state caused the effect to fire.
+
+```svelte
+<script>
+	import { doSomeWork } from './elsewhere';
+
+	$effect(() => {
+		+++// $inspect.trace must be the first statement of a function body+++
+		+++$inspect.trace();+++
+		doSomeWork();
+	});
+</script>
+```
+
+`$inspect.trace` takes an optional first argument which will be used as the label.

packages/opencode/skills/svelte-core-bestpractices/references/@attach.md

@@ -0,0 +1,166 @@
+Attachments are functions that run in an [effect]($effect) when an element is mounted to the DOM or when [state]($state) read inside the function updates.
+
+Optionally, they can return a function that is called before the attachment re-runs, or after the element is later removed from the DOM.
+
+> [!NOTE]
+> Attachments are available in Svelte 5.29 and newer.
+
+```svelte
+<!--- file: App.svelte --->
+<script>
+	/** @type {import('svelte/attachments').Attachment} */
+	function myAttachment(element) {
+		console.log(element.nodeName); // 'DIV'
+
+		return () => {
+			console.log('cleaning up');
+		};
+	}
+</script>
+
+<div {@attach myAttachment}>...</div>
+```
+
+An element can have any number of attachments.
+
+## Attachment factories
+
+A useful pattern is for a function, such as `tooltip` in this example, to _return_ an attachment (demo:
+
+```svelte
+<!--- file: App.svelte --->
+<script>
+	import tippy from 'tippy.js';
+
+	let content = $state('Hello!');
+
+	/**
+	 * @param {string} content
+	 * @returns {import('svelte/attachments').Attachment}
+	 */
+	function tooltip(content) {
+		return (element) => {
+			const tooltip = tippy(element, { content });
+			return tooltip.destroy;
+		};
+	}
+</script>
+
+<input bind:value={content} />
+
+<button {@attach tooltip(content)}> Hover me </button>
+```
+
+Since the `tooltip(content)` expression runs inside an [effect]($effect), the attachment will be destroyed and recreated whenever `content` changes. The same thing would happen for any state read _inside_ the attachment function when it first runs. (If this isn't what you want, see [Controlling when attachments re-run](#Controlling-when-attachments-re-run).)
+
+## Inline attachments
+
+Attachments can also be created inline (demo:
+
+```svelte
+<!--- file: App.svelte --->
+<canvas
+	width={32}
+	height={32}
+	{@attach (canvas) => {
+		const context = canvas.getContext('2d');
+
+		$effect(() => {
+			context.fillStyle = color;
+			context.fillRect(0, 0, canvas.width, canvas.height);
+		});
+	}}
+></canvas>
+```
+
+> [!NOTE]
+> The nested effect runs whenever `color` changes, while the outer effect (where `canvas.getContext(...)` is called) only runs once, since it doesn't read any reactive state.
+
+## Conditional attachments
+
+Falsy values like `false` or `undefined` are treated as no attachment, enabling conditional usage:
+
+```svelte
+<div {@attach enabled && myAttachment}>...</div>
+```
+
+## Passing attachments to components
+
+When used on a component, `{@attach ...}` will create a prop whose key is a [`Symbol`](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Symbol). If the component then [spreads](/tutorial/svelte/spread-props) props onto an element, the element will receive those attachments.
+
+This allows you to create _wrapper components_ that augment elements (demo:
+
+```svelte
+<!--- file: Button.svelte --->
+<script>
+	/** @type {import('svelte/elements').HTMLButtonAttributes} */
+	let { children, ...props } = $props();
+</script>
+
+<!-- `props` includes attachments -->
+<button {...props}>
+	{@render children?.()}
+</button>
+```
+
+```svelte
+<!--- file: App.svelte --->
+<script>
+	import tippy from 'tippy.js';
+	import Button from './Button.svelte';
+
+	let content = $state('Hello!');
+
+	/**
+	 * @param {string} content
+	 * @returns {import('svelte/attachments').Attachment}
+	 */
+	function tooltip(content) {
+		return (element) => {
+			const tooltip = tippy(element, { content });
+			return tooltip.destroy;
+		};
+	}
+</script>
+
+<input bind:value={content} />
+
+<Button {@attach tooltip(content)}>Hover me</Button>
+```
+
+## Controlling when attachments re-run
+
+Attachments, unlike [actions](use), are fully reactive: `{@attach foo(bar)}` will re-run on changes to `foo` _or_ `bar` (or any state read inside `foo`):
+
+```js
+// @errors: 7006 2304 2552
+function foo(bar) {
+	return (node) => {
+		veryExpensiveSetupWork(node);
+		update(node, bar);
+	};
+}
+```
+
+In the rare case that this is a problem (for example, if `foo` does expensive and unavoidable setup work) consider passing the data inside a function and reading it in a child effect:
+
+```js
+// @errors: 7006 2304 2552
+function foo(+++getBar+++) {
+	return (node) => {
+		veryExpensiveSetupWork(node);
+
++++		$effect(() => {
+			update(node, getBar());
+		});+++
+	}
+}
+```
+
+## Creating attachments programmatically
+
+To add attachments to an object that will be spread onto a component or element, use [`createAttachmentKey`](svelte-attachments#createAttachmentKey).
+
+## Converting actions to attachments
+
+If you're using a library that only provides actions, you can convert them to attachments with [`fromAction`](svelte-attachments#fromAction), allowing you to (for example) use them with components.

packages/opencode/skills/svelte-core-bestpractices/references/@render.md

@@ -0,0 +1,35 @@
+To render a [snippet](snippet), use a `{@render ...}` tag.
+
+```svelte
+{#snippet sum(a, b)}
+	<p>{a} + {b} = {a + b}</p>
+{/snippet}
+
+{@render sum(1, 2)}
+{@render sum(3, 4)}
+{@render sum(5, 6)}
+```
+
+The expression can be an identifier like `sum`, or an arbitrary JavaScript expression:
+
+```svelte
+{@render (cool ? coolSnippet : lameSnippet)()}
+```
+
+## Optional snippets
+
+If the snippet is potentially undefined — for example, because it's an incoming prop — then you can use optional chaining to only render it when it _is_ defined:
+
+```svelte
+{@render children?.()}
+```
+
+Alternatively, use an [`{#if ...}`](if) block with an `:else` clause to render fallback content:
+
+```svelte
+{#if children}
+	{@render children()}
+{:else}
+	<p>fallback content</p>
+{/if}
+```