feat: add meeting message shortcut

buildListParams used to re-call resolveFolderID / resolveFolderName (and the
label counterparts) on every list page to assemble folder_id / label_id.
Because resolveListFilter already resolves the filter once before the
pagination loop, the second pass hit the folders/labels list API again on
every page — 1 + page_count calls total, which easily trips rate limits.

buildListParams now only assembles API params from the already-resolved
FolderID / LabelID produced by resolveListFilter; it no longer resolves
names or aliases. The default folder_id=INBOX is still applied when no
explicit filter is present, and only overridden when the caller supplied a
canonical folder ID. The runtime / mailboxID / dryRun parameters are kept
for signature stability (resolveListFilter and buildSearchParams share the
same call shape).

Adds TestMailTriageCustomFolderResolvesOnceAcrossListPages: a custom-folder
filter forced across two messages-list pages, with a non-reusable folders
list stub so any second folders API call fails the test. Updated the two
existing buildListParams alias tests to run resolveListFilter first, mirroring
the real DryRun/Execute call order.

sprint: S1

Co-authored-by: xukuncx <283114605+xukuncx@users.noreply.github.com>

.github/workflows/semantic-review.yml

@@ -47,10 +47,13 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            let eventBaseSha = runPRs[0]?.base?.sha || "";
+            const eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = eventHeadSha || run.head_sha;
+            const targetHeadSha = run.head_sha;
             if (!/^[a-f0-9]{40}$/i.test(targetHeadSha)) throw new Error("invalid PR head sha");
+            if (eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
+              core.notice("PR quality summary using workflow_run head_sha because workflow_run pull request head differs from the CI run head");
+            }
 
             const factsArtifactPattern = /^quality-gate-facts-([a-f0-9]{40})-([a-f0-9]{40})$/i;
             const { data: artifactData } = await github.rest.actions.listWorkflowRunArtifacts({
@@ -71,11 +74,11 @@ jobs:
               if (artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
                 artifactError = "facts artifact head sha does not match verified PR head sha";
                 factsArtifactName = "";
-              } else if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
-                artifactError = "facts artifact base sha does not match workflow_run pull request base sha";
-                factsArtifactName = "";
               } else {
                 artifactBaseSha = parsedBaseSha;
+                if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
+                  core.notice("PR quality summary using facts artifact base because workflow_run pull request base differs from the CI facts artifact base");
+                }
               }
             }
             if (!prNumber) {
@@ -123,7 +126,7 @@ jobs:
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
+            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/i.test(baseSha)) throw new Error("invalid PR base sha");
             if ((eventBaseSha || artifactBaseSha) && pr.base.sha !== baseSha) {
               core.notice("PR quality summary skipped: workflow_run is stale for this PR base");
@@ -255,10 +258,13 @@ jobs:
               throw new Error(`ambiguous workflow_run pull request bindings: ${runPRs.length}`);
             }
             let prNumber = Number(runPRs[0]?.number || 0);
-            let eventBaseSha = runPRs[0]?.base?.sha || "";
+            const eventBaseSha = runPRs[0]?.base?.sha || "";
             const eventHeadSha = runPRs[0]?.head?.sha || "";
-            const targetHeadSha = eventHeadSha || run.head_sha;
+            const targetHeadSha = run.head_sha;
             if (!/^[a-f0-9]{40}$/i.test(targetHeadSha)) throw new Error("invalid PR head sha");
+            if (eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
+              core.notice("semantic review using workflow_run head_sha because workflow_run pull request head differs from the CI run head");
+            }
 
             const factsArtifactPattern = /^quality-gate-facts-([a-f0-9]{40})-([a-f0-9]{40})$/i;
             const { data: artifactData } = await github.rest.actions.listWorkflowRunArtifacts({
@@ -279,11 +285,11 @@ jobs:
               if (artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()) {
                 artifactError = "facts artifact head sha does not match verified PR head sha";
                 factsArtifactName = "";
-              } else if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
-                artifactError = "facts artifact base sha does not match workflow_run pull request base sha";
-                factsArtifactName = "";
               } else {
                 artifactBaseSha = parsedBaseSha;
+                if (eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()) {
+                  core.notice("semantic review using facts artifact base because workflow_run pull request base differs from the CI facts artifact base");
+                }
               }
             }
             if (!prNumber) {
@@ -331,7 +337,7 @@ jobs:
               core.setOutput("stale", "true");
               return;
             }
-            const baseSha = eventBaseSha || artifactBaseSha || pr.base.sha;
+            const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha;
             if (!/^[a-f0-9]{40}$/i.test(baseSha)) throw new Error("invalid PR base sha");
             if ((eventBaseSha || artifactBaseSha) && pr.base.sha !== baseSha) {
               core.notice("semantic review skipped: workflow_run is stale for this PR base");

scripts/semantic-review-workflow.test.sh

@@ -179,7 +179,10 @@ fi
 require_in_step "$summary_verify_step" 'workflowPath !== ".github/workflows/ci.yml"' "PR quality summary must verify the triggering workflow path"
 require_in_step "$summary_verify_step" 'run.event !== "pull_request"' "PR quality summary must only handle pull_request workflow_run events"
 require_in_step "$summary_verify_step" 'run.repository.id !== context.payload.repository.id' "PR quality summary must verify workflow_run repository id"
+require_in_step "$summary_verify_step" 'const targetHeadSha = run.head_sha' "PR quality summary must use the CI run head SHA as the verified PR head"
+require_in_step "$summary_verify_step" 'eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "PR quality summary should tolerate mutable workflow_run PR head metadata"
 require_in_step "$summary_verify_step" 'factsArtifactPattern' "PR quality summary should use the base-bound facts artifact name when available"
+require_in_step "$summary_verify_step" 'const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha' "PR quality summary must prefer the CI-time artifact base SHA"
 require_in_step "$summary_verify_step" 'core.setOutput("artifact_error"' "PR quality summary must expose artifact binding failures"
 require_in_step "$summary_artifact_step" 'factsArtifactName' "PR quality summary artifact step must use the verified facts artifact binding"
 require_in_step "$summary_extract_facts_step" 'SEMANTIC_REVIEW_DECISION_OUT' "PR quality summary artifact verifier must write an infrastructure decision on verifier failure"
@@ -198,8 +201,9 @@ require_in_step "$verify_step" 'workflowPath !== ".github/workflows/ci.yml"' "se
 require_in_step "$verify_step" 'run.repository.id !== context.payload.repository.id' "semantic-review must verify workflow_run repository id"
 require_in_step "$verify_step" 'run.event !== "pull_request"' "semantic-review must only handle pull_request workflow_run events"
 require_in_step "$verify_step" 'run.conclusion !== "success"' "semantic-review must only consume successful CI runs"
-require_in_step "$verify_step" 'const eventHeadSha = runPRs[0]?.head?.sha || ""' "semantic-review must prefer workflow_run PR head when GitHub provides it"
-require_in_step "$verify_step" 'const targetHeadSha = eventHeadSha || run.head_sha' "semantic-review target PR head must come from the workflow_run event"
+require_in_step "$verify_step" 'const eventHeadSha = runPRs[0]?.head?.sha || ""' "semantic-review should inspect workflow_run PR head metadata"
+require_in_step "$verify_step" 'const targetHeadSha = run.head_sha' "semantic-review target PR head must come from the completed CI run"
+require_in_step "$verify_step" 'eventHeadSha && eventHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "semantic-review should tolerate mutable workflow_run PR head metadata"
 require_in_step "$verify_step" 'factsArtifactPattern' "semantic-review must use a base-bound facts artifact name"
 require_in_step "$verify_step" 'listWorkflowRunArtifacts' "semantic-review must read the workflow_run artifacts before resolving fallback base SHA"
 require_in_step "$verify_step" 'artifactHeadSha.toLowerCase() !== targetHeadSha.toLowerCase()' "semantic-review must not let the artifact choose a different PR head"
@@ -210,8 +214,8 @@ require_in_step "$verify_step" 'commit_sha: targetHeadSha' "semantic-review fall
 require_in_step "$verify_step" 'github.rest.pulls.list' "semantic-review must have a pull-list fallback when commit association is empty"
 require_in_step "$verify_step" 'candidatePRs.length > 1' "semantic-review must fail closed when commit-to-PR fallback is ambiguous"
 require_in_step "$verify_step" 'pr.head.sha !== targetHeadSha' "semantic-review must skip stale PR heads"
-require_in_step "$verify_step" 'eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()' "semantic-review must reject mismatched event and artifact base SHAs"
-require_in_step "$verify_step" 'const baseSha = eventBaseSha || artifactBaseSha' "semantic-review fallback must use the CI-time artifact base SHA"
+require_in_step "$verify_step" 'eventBaseSha && parsedBaseSha.toLowerCase() !== eventBaseSha.toLowerCase()' "semantic-review should tolerate mutable workflow_run PR base metadata"
+require_in_step "$verify_step" 'const baseSha = artifactBaseSha || eventBaseSha || pr.base.sha' "semantic-review must prefer the CI-time artifact base SHA"
 require_in_step "$verify_step" 'pr.base.sha !== baseSha' "semantic-review must skip stale PR bases"
 require_in_step "$verify_step" 'core.setOutput("run_id"' "semantic-review must pass verified workflow run id to publisher"
 require_in_step "$verify_step" 'core.setOutput("head_repo_id"' "semantic-review must pass verified head repo id"

shortcuts/mail/mail_triage.go

@@ -770,13 +770,7 @@ func buildListParams(runtime *common.RuntimeContext, mailboxID string, f triageF
 				params["folder_id"] = folderIDFromFilter
 			}
 		} else {
-			resolved, err := resolveFolderID(runtime, mailboxID, folderIDFromFilter)
-			if err != nil {
-				return nil, err
-			}
-			if resolved != "" {
-				params["folder_id"] = resolved
-			}
+			params["folder_id"] = folderIDFromFilter
 		}
 	} else if folderFromFilter != "" {
 		if dryRun {
@@ -786,13 +780,7 @@ func buildListParams(runtime *common.RuntimeContext, mailboxID string, f triageF
 				params["folder_id"] = folderFromFilter
 			}
 		} else {
-			resolved, err := resolveFolderName(runtime, mailboxID, folderFromFilter)
-			if err != nil {
-				return nil, err
-			}
-			if resolved != "" {
-				params["folder_id"] = resolved
-			}
+			params["folder_id"] = folderFromFilter
 		}
 	}
 
@@ -811,13 +799,7 @@ func buildListParams(runtime *common.RuntimeContext, mailboxID string, f triageF
 				params["label_id"] = labelIDFromFilter
 			}
 		} else {
-			resolved, err := resolveLabelID(runtime, mailboxID, labelIDFromFilter)
-			if err != nil {
-				return nil, err
-			}
-			if resolved != "" {
-				params["label_id"] = resolved
-			}
+			params["label_id"] = labelIDFromFilter
 		}
 	} else if labelFromFilter != "" {
 		if dryRun {
@@ -827,13 +809,7 @@ func buildListParams(runtime *common.RuntimeContext, mailboxID string, f triageF
 				params["label_id"] = labelFromFilter
 			}
 		} else {
-			resolved, err := resolveLabelName(runtime, mailboxID, labelFromFilter)
-			if err != nil {
-				return nil, err
-			}
-			if resolved != "" {
-				params["label_id"] = resolved
-			}
+			params["label_id"] = labelFromFilter
 		}
 	}
 

shortcuts/mail/mail_triage_test.go

@@ -12,6 +12,7 @@ import (
 	"testing"
 
 	"github.com/larksuite/cli/errs"
+	"github.com/larksuite/cli/internal/auth"
 	"github.com/larksuite/cli/internal/cmdutil"
 	"github.com/larksuite/cli/internal/httpmock"
 	"github.com/larksuite/cli/shortcuts/common"
@@ -974,7 +975,11 @@ func TestBuildListParamsDryRunOnlyUnread(t *testing.T) {
 func TestBuildListParamsDryRunFolderAlias(t *testing.T) {
 	rt := runtimeForMailTriageTest(t, nil)
 	f := triageFilter{Folder: "sent"}
-	got, err := buildListParams(rt, "me", f, 20, "", true)
+	resolved, err := resolveListFilter(rt, "me", f, true)
+	if err != nil {
+		t.Fatalf("resolveListFilter: %v", err)
+	}
+	got, err := buildListParams(rt, "me", resolved, 20, "", true)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -983,10 +988,30 @@ func TestBuildListParamsDryRunFolderAlias(t *testing.T) {
 	}
 }
 
+func TestBuildListParamsDryRunCustomFolderPreservesInput(t *testing.T) {
+	rt := runtimeForMailTriageTest(t, nil)
+	f := triageFilter{Folder: "team-folder"}
+	resolved, err := resolveListFilter(rt, "me", f, true)
+	if err != nil {
+		t.Fatalf("resolveListFilter: %v", err)
+	}
+	got, err := buildListParams(rt, "me", resolved, 20, "", true)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if got["folder_id"] != "team-folder" {
+		t.Fatalf("expected dry-run folder_id=team-folder, got %v", got["folder_id"])
+	}
+}
+
 func TestBuildListParamsDryRunLabelAlias(t *testing.T) {
 	rt := runtimeForMailTriageTest(t, nil)
 	f := triageFilter{Label: "flagged"}
-	got, err := buildListParams(rt, "me", f, 10, "", true)
+	resolved, err := resolveListFilter(rt, "me", f, true)
+	if err != nil {
+		t.Fatalf("resolveListFilter: %v", err)
+	}
+	got, err := buildListParams(rt, "me", resolved, 10, "", true)
 	if err != nil {
 		t.Fatal(err)
 	}
@@ -995,6 +1020,25 @@ func TestBuildListParamsDryRunLabelAlias(t *testing.T) {
 	}
 }
 
+func TestBuildListParamsDryRunCustomLabelPreservesInput(t *testing.T) {
+	rt := runtimeForMailTriageTest(t, nil)
+	f := triageFilter{Label: "custom-label"}
+	resolved, err := resolveListFilter(rt, "me", f, true)
+	if err != nil {
+		t.Fatalf("resolveListFilter: %v", err)
+	}
+	got, err := buildListParams(rt, "me", resolved, 10, "", true)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, ok := got["folder_id"]; ok {
+		t.Fatalf("folder_id should not be set when label is specified, got %v", got["folder_id"])
+	}
+	if got["label_id"] != "custom-label" {
+		t.Fatalf("expected dry-run label_id=custom-label, got %v", got["label_id"])
+	}
+}
+
 // --- buildSearchParams additional coverage ---
 
 func TestBuildSearchParamsAllFilterFields(t *testing.T) {
@@ -1791,3 +1835,137 @@ func mailTriageSearchItem(messageID, subject string) map[string]interface{} {
 		},
 	}
 }
+
+// registerMailTriageFoldersListStub registers a NON-reusable stub for the
+// mailbox folders list API. Because it is non-reusable, any second hit returns
+// "httpmock: no stub for GET .../folders" — which is exactly the assertion we
+// use to prove resolveListFilter runs once and buildListParams does NOT
+// re-resolve. folderID/folderName is the single custom folder the API reports.
+func registerMailTriageFoldersListStub(reg *httpmock.Registry, mailbox, folderID, folderName string) {
+	reg.Register(&httpmock.Stub{
+		Method: "GET",
+		URL:    mailboxPath(mailbox, "folders"),
+		Body: map[string]interface{}{
+			"code": 0,
+			"data": map[string]interface{}{
+				"items": []interface{}{
+					map[string]interface{}{
+						"id":   folderID,
+						"name": folderName,
+					},
+				},
+			},
+		},
+	})
+}
+
+// registerMailTriageListPageStub registers one page of the messages list API,
+// disambiguated from sibling pages by a URL substring unique to that page
+// (e.g. "page_size=5" for page 1 vs "page_size=2" for page 2). The substring
+// must NOT depend on query-param ordering: map iteration makes param order
+// nondeterministic, so prefer a value-only token like "page_size=N" (the N
+// differs per page because pageSize = maxCount - fetched_so_far). Non-reusable
+// so reg.Verify catches under- or over-consumption.
+func registerMailTriageListPageStub(reg *httpmock.Registry, urlSubstring string, items []string, hasMore bool, pageToken string) {
+	data := map[string]interface{}{
+		"items":    items,
+		"has_more": hasMore,
+	}
+	if pageToken != "" {
+		data["page_token"] = pageToken
+	}
+	reg.Register(&httpmock.Stub{
+		Method: "GET",
+		URL:    urlSubstring,
+		Body: map[string]interface{}{
+			"code": 0,
+			"data": data,
+		},
+	})
+}
+
+// TestMailTriageCustomFolderResolvesOnceAcrossListPages is the regression test
+// for the bug where buildListParams re-called resolveFolderID on every list
+// page, turning "resolve once" into "1 + page_count" folder-list API calls and
+// easily tripping rate limits.
+//
+// Setup: a custom folder filter that forces resolveListFilter to hit the
+// folders list API once (to map folder name "team-folder" to folder_id), then two
+// messages-list pages. The folders list stub is non-reusable, so if
+// buildListParams re-resolves, the second hit fails with "no stub". The
+// messages-list stubs are page-specific (disambiguated by page_size in the
+// URL), so both pages are served and Verify asserts each fired exactly once.
+func TestMailTriageCustomFolderResolvesOnceAcrossListPages(t *testing.T) {
+	f, stdout, _, reg := mailShortcutTestFactory(t)
+	defer reg.Verify(t)
+
+	// listMailboxFolders (called once by resolveListFilter) gates on the
+	// mail:user_mailbox.folder:read scope, which the default test token does
+	// not carry. Re-store the token with that scope appended so the folders
+	// API call is actually exercised (and thus the non-reusable folders stub
+	// is the load-bearing "exactly once" assertion).
+	const folderScope = "mail:user_mailbox.folder:read"
+	cfg := mailTestConfig()
+	if stored := auth.GetStoredToken(cfg.AppID, cfg.UserOpenId); stored != nil {
+		if !strings.Contains(stored.Scope, folderScope) {
+			stored.Scope = stored.Scope + " " + folderScope
+			if err := auth.SetStoredToken(stored); err != nil {
+				t.Fatalf("re-store token with folder scope: %v", err)
+			}
+		}
+	}
+
+	const (
+		mailbox    = "me"
+		folderName = "team-folder"
+		folderID   = "fld_custom_team"
+		page2Token = "tok_page2"
+	)
+	// --max 5 with listPageMax=20 → pageSize = 5-0 = 5 on page 1, then 5-3 = 2
+	// on page 2. The page_size query value disambiguates the two list stubs.
+	page1IDs := []string{"msg_a", "msg_b", "msg_c"}
+	page2IDs := []string{"msg_d", "msg_e"}
+
+	// Folders list: registered exactly once, non-reusable. Any second folder
+	// lookup (the bug) fails the test with "no stub for GET .../folders".
+	registerMailTriageFoldersListStub(reg, mailbox, folderID, folderName)
+	// Messages list, page 1: 3 ids, has_more, hands off a page-2 token. The
+	// page_size value (5 = maxCount - 0) is unique to page 1; page 2 uses 2.
+	registerMailTriageListPageStub(reg, "page_size=5", page1IDs, true, page2Token)
+	// Messages list, page 2: 2 ids, terminal.
+	registerMailTriageListPageStub(reg, "page_size=2", page2IDs, false, "")
+	// Batch metadata fetch for all 5 ids.
+	registerMailTriageBatchStub(reg, mailbox, []map[string]interface{}{
+		mailTriageBatchMessage("msg_a", "Subject A"),
+		mailTriageBatchMessage("msg_b", "Subject B"),
+		mailTriageBatchMessage("msg_c", "Subject C"),
+		mailTriageBatchMessage("msg_d", "Subject D"),
+		mailTriageBatchMessage("msg_e", "Subject E"),
+	})
+
+	args := []string{
+		"+triage",
+		"--as", "user",
+		"--mailbox", mailbox,
+		"--filter", `{"folder":"` + folderName + `"}`,
+		"--max", "5",
+		"--format", "json",
+	}
+	if err := runMountedMailShortcut(t, MailTriage, args, f, stdout); err != nil {
+		t.Fatalf("unexpected error running +triage (likely a second folders API call — the bug): %v", err)
+	}
+
+	data := decodeMailTriageJSONOutput(t, stdout)
+	messages := mailTriageMessagesFromOutput(t, data)
+	if len(messages) != 5 {
+		t.Fatalf("expected 5 messages across 2 pages, got %d (stdout=%s)", len(messages), stdout.String())
+	}
+	if got := data["has_more"]; got != false {
+		t.Fatalf("expected has_more=false after exhausting pages, got %v", got)
+	}
+	// All registered stubs (1 folders + 2 list pages + 1 batch_get) are
+	// non-reusable; reg.Verify (deferred above) asserts each was matched
+	// exactly once. Combined with the non-reusable folders stub, this is the
+	// proof that the folders list API was called exactly once across both
+	// pages — the core invariant the fix restores.
+}

shortcuts/vc/shortcuts.go

@@ -15,5 +15,6 @@ func Shortcuts() []common.Shortcut {
 		VCMeetingLeave,
 		VCMeetingListActive,
 		VCMeetingEvents,
+		VCMeetingMessageSend,
 	}
 }

shortcuts/vc/vc_meeting_events_test.go

@@ -838,7 +838,7 @@ func TestVCShortcuts_RegistersMeetingAgentCommands(t *testing.T) {
 	for _, shortcut := range got {
 		commands = append(commands, shortcut.Command)
 	}
-	want := []string{"+search", "+notes", "+recording", "+meeting-join", "+meeting-leave", "+meeting-list-active", "+meeting-events"}
+	want := []string{"+search", "+notes", "+recording", "+meeting-join", "+meeting-leave", "+meeting-list-active", "+meeting-events", "+meeting-message-send"}
 	if !reflect.DeepEqual(commands, want) {
 		t.Fatalf("shortcut commands = %#v, want %#v", commands, want)
 	}

shortcuts/vc/vc_meeting_message_send.go

@@ -0,0 +1,144 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package vc
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/larksuite/cli/errs"
+	"github.com/larksuite/cli/shortcuts/common"
+)
+
+const (
+	meetingMessageTypeText     = "text"
+	meetingMessageTypeReaction = "reaction"
+)
+
+var supportedVCMeetingReactionKeys = map[string]struct{}{
+	"VC_CanNotSee":   {},
+	"VC_NoSound":     {},
+	"VC_LooksGood":   {},
+	"VC_SoundsClear": {},
+}
+
+// VCMeetingMessageSend sends an in-meeting text message or feedback reaction.
+var VCMeetingMessageSend = common.Shortcut{
+	Service:     "vc",
+	Command:     "+meeting-message-send",
+	Description: "Send an in-meeting text message or feedback reaction",
+	Risk:        "write",
+	Scopes:      []string{"vc:meeting.message:write"},
+	AuthTypes:   []string{"user", "bot"},
+	HasFormat:   true,
+	Flags: []common.Flag{
+		{Name: "meeting-id", Required: true, Desc: "meeting ID to send into"},
+		{Name: "msg-type", Desc: "message type: text or reaction"},
+		{Name: "text", Desc: "text content when --msg-type text"},
+		{Name: "emoji-type", Desc: "reaction key when --msg-type reaction"},
+		{Name: "uuid", Desc: "optional idempotency key"},
+	},
+	Validate: func(ctx context.Context, runtime *common.RuntimeContext) error {
+		if err := validateMeetingEventsMeetingID(runtime.Str("meeting-id")); err != nil {
+			return err
+		}
+		_, err := resolveMeetingMessageType(runtime)
+		return err
+	},
+	DryRun: func(ctx context.Context, runtime *common.RuntimeContext) *common.DryRunAPI {
+		body, err := buildMeetingMessageSendBody(runtime)
+		if err != nil {
+			return common.NewDryRunAPI().Set("error", err.Error())
+		}
+		return common.NewDryRunAPI().
+			POST(buildMeetingMessageSendPath(runtime.Str("meeting-id"))).
+			Body(body)
+	},
+	Execute: func(ctx context.Context, runtime *common.RuntimeContext) error {
+		body, err := buildMeetingMessageSendBody(runtime)
+		if err != nil {
+			return err
+		}
+		data, err := runtime.CallAPITyped(http.MethodPost, buildMeetingMessageSendPath(runtime.Str("meeting-id")), nil, body)
+		if err != nil {
+			return err
+		}
+		if data == nil {
+			data = map[string]interface{}{}
+		}
+		runtime.OutFormat(data, nil, func(w io.Writer) {
+			fmt.Fprintln(w, "Meeting message sent.")
+			if msgType := common.GetString(data, "msg_type"); msgType != "" {
+				fmt.Fprintf(w, "  Type:  %s\n", msgType)
+			} else if msgType, _ := body["msg_type"].(string); msgType != "" {
+				fmt.Fprintf(w, "  Type:  %s\n", msgType)
+			}
+			if uuid := common.GetString(data, "uuid"); uuid != "" {
+				fmt.Fprintf(w, "  UUID:  %s\n", uuid)
+			}
+		})
+		return nil
+	},
+}
+
+func buildMeetingMessageSendPath(meetingID string) string {
+	return fmt.Sprintf("/open-apis/vc/v1/meetings/%s/messages", strings.TrimSpace(meetingID))
+}
+
+func buildMeetingMessageSendBody(runtime *common.RuntimeContext) (map[string]interface{}, error) {
+	msgType, err := resolveMeetingMessageType(runtime)
+	if err != nil {
+		return nil, err
+	}
+	body := map[string]interface{}{
+		"msg_type": msgType,
+	}
+	switch msgType {
+	case meetingMessageTypeText:
+		body["text"] = strings.TrimSpace(runtime.Str("text"))
+	case meetingMessageTypeReaction:
+		body["emoji_type"] = strings.TrimSpace(runtime.Str("emoji-type"))
+	}
+	if uuid := strings.TrimSpace(runtime.Str("uuid")); uuid != "" {
+		body["uuid"] = uuid
+	}
+	return body, nil
+}
+
+func resolveMeetingMessageType(runtime *common.RuntimeContext) (string, error) {
+	msgType := strings.ToLower(strings.TrimSpace(runtime.Str("msg-type")))
+	text := strings.TrimSpace(runtime.Str("text"))
+	emojiType := strings.TrimSpace(runtime.Str("emoji-type"))
+
+	if msgType == "" {
+		switch {
+		case text != "" && emojiType == "":
+			msgType = meetingMessageTypeText
+		case text == "" && emojiType != "":
+			msgType = meetingMessageTypeReaction
+		default:
+			return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "--msg-type is required when both --text and --emoji-type are empty or both are set").WithParam("--msg-type")
+		}
+	}
+
+	switch msgType {
+	case meetingMessageTypeText:
+		if text == "" {
+			return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "--text is required when --msg-type text").WithParam("--text")
+		}
+	case meetingMessageTypeReaction:
+		if emojiType == "" {
+			return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "--emoji-type is required when --msg-type reaction").WithParam("--emoji-type")
+		}
+		if _, ok := supportedVCMeetingReactionKeys[emojiType]; !ok {
+			return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "--emoji-type must be one of VC_CanNotSee, VC_NoSound, VC_LooksGood, VC_SoundsClear").WithParam("--emoji-type")
+		}
+	default:
+		return "", errs.NewValidationError(errs.SubtypeInvalidArgument, "--msg-type must be text or reaction").WithParam("--msg-type")
+	}
+	return msgType, nil
+}

shortcuts/vc/vc_meeting_message_send_test.go

@@ -0,0 +1,125 @@
+// Copyright (c) 2026 Lark Technologies Pte. Ltd.
+// SPDX-License-Identifier: MIT
+
+package vc
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/spf13/cobra"
+
+	"github.com/larksuite/cli/internal/cmdutil"
+	"github.com/larksuite/cli/shortcuts/common"
+)
+
+func newMeetingMessageSendRuntime() *common.RuntimeContext {
+	cmd := &cobra.Command{Use: "test"}
+	cmd.Flags().String("meeting-id", "", "")
+	cmd.Flags().String("msg-type", "", "")
+	cmd.Flags().String("text", "", "")
+	cmd.Flags().String("emoji-type", "", "")
+	cmd.Flags().String("uuid", "", "")
+	return common.TestNewRuntimeContext(cmd, defaultConfig())
+}
+
+func mustSetMeetingMessageSendFlag(t *testing.T, runtime *common.RuntimeContext, name, value string) {
+	t.Helper()
+	if err := runtime.Cmd.Flags().Set(name, value); err != nil {
+		t.Fatalf("Flags().Set(%q, %q) error = %v", name, value, err)
+	}
+}
+
+func TestMeetingMessageSendBuildBody_Text(t *testing.T) {
+	runtime := newMeetingMessageSendRuntime()
+	mustSetMeetingMessageSendFlag(t, runtime, "text", " hello ")
+	mustSetMeetingMessageSendFlag(t, runtime, "uuid", " cid-1 ")
+
+	body, err := buildMeetingMessageSendBody(runtime)
+	if err != nil {
+		t.Fatalf("buildMeetingMessageSendBody() error = %v", err)
+	}
+	if body["msg_type"] != meetingMessageTypeText {
+		t.Fatalf("msg_type = %v, want text", body["msg_type"])
+	}
+	if body["text"] != "hello" {
+		t.Fatalf("text = %v, want hello", body["text"])
+	}
+	if body["uuid"] != "cid-1" {
+		t.Fatalf("uuid = %v, want cid-1", body["uuid"])
+	}
+}
+
+func TestMeetingMessageSendBuildBody_Reaction(t *testing.T) {
+	runtime := newMeetingMessageSendRuntime()
+	mustSetMeetingMessageSendFlag(t, runtime, "msg-type", "reaction")
+	mustSetMeetingMessageSendFlag(t, runtime, "emoji-type", "VC_LooksGood")
+
+	body, err := buildMeetingMessageSendBody(runtime)
+	if err != nil {
+		t.Fatalf("buildMeetingMessageSendBody() error = %v", err)
+	}
+	if body["msg_type"] != meetingMessageTypeReaction {
+		t.Fatalf("msg_type = %v, want reaction", body["msg_type"])
+	}
+	if body["emoji_type"] != "VC_LooksGood" {
+		t.Fatalf("emoji_type = %v, want VC_LooksGood", body["emoji_type"])
+	}
+	if _, ok := body["text"]; ok {
+		t.Fatalf("text should be omitted for reaction, got %#v", body["text"])
+	}
+}
+
+func TestMeetingMessageSendValidateRejectsMeetingNumber(t *testing.T) {
+	runtime := newMeetingMessageSendRuntime()
+	mustSetMeetingMessageSendFlag(t, runtime, "meeting-id", "123456789")
+	mustSetMeetingMessageSendFlag(t, runtime, "text", "hello")
+
+	err := VCMeetingMessageSend.Validate(context.Background(), runtime)
+	if err == nil {
+		t.Fatal("expected validation error")
+	}
+	if !strings.Contains(err.Error(), "9-digit meeting number") {
+		t.Fatalf("error = %v, want 9-digit meeting number hint", err)
+	}
+}
+
+func TestMeetingMessageSendValidateRejectsUnknownReaction(t *testing.T) {
+	runtime := newMeetingMessageSendRuntime()
+	mustSetMeetingMessageSendFlag(t, runtime, "meeting-id", "7651377260537433044")
+	mustSetMeetingMessageSendFlag(t, runtime, "msg-type", "reaction")
+	mustSetMeetingMessageSendFlag(t, runtime, "emoji-type", "LOVE")
+
+	err := VCMeetingMessageSend.Validate(context.Background(), runtime)
+	if err == nil {
+		t.Fatal("expected validation error")
+	}
+	if !strings.Contains(err.Error(), "VC_CanNotSee") {
+		t.Fatalf("error = %v, want supported key hint", err)
+	}
+}
+
+func TestMeetingMessageSendDryRun_Text(t *testing.T) {
+	f, stdout, _, _ := cmdutil.TestFactory(t, defaultConfig())
+	err := mountAndRun(t, VCMeetingMessageSend, []string{
+		"+meeting-message-send", "--dry-run", "--as", "user",
+		"--meeting-id", "7651377260537433044",
+		"--text", "hello",
+		"--uuid", "cid-1",
+	}, f, stdout)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	out := stdout.String()
+	for _, want := range []string{
+		"/open-apis/vc/v1/meetings/7651377260537433044/messages",
+		"\"msg_type\": \"text\"",
+		"\"text\": \"hello\"",
+		"\"uuid\": \"cid-1\"",
+	} {
+		if !strings.Contains(out, want) {
+			t.Fatalf("dry-run output missing %q: %s", want, out)
+		}
+	}
+}

skills/lark-doc/SKILL.md

@@ -26,7 +26,7 @@ lark-cli docs +update --api-version v2 --doc "文档URL或token" --command appen
 **CRITICAL — 执行对应操作前，MUST 先用 Read 工具读取以下文件，缺一不可：**
 1. [`../lark-shared/SKILL.md`](../lark-shared/SKILL.md) — 认证、权限处理、全局参数（所有操作通用）
 2. **读取文档（`docs +fetch --api-version v2`）** → 必读 [`lark-doc-fetch.md`](references/lark-doc-fetch.md)（`--scope` / `--detail` 选择、局部读取策略、`<fragment>` / `<excerpt>` 输出结构）
-3. **创建或编辑文档内容** → 必读 [`lark-doc-xml.md`](references/lark-doc-xml.md)（XML 语法规则，仅当用户明确要求 Markdown 时改读 [`lark-doc-md.md`](references/lark-doc-md.md)）；从零创建时加读 [`lark-doc-create-workflow.md`](references/style/lark-doc-create-workflow.md)；编辑已有文档时加读 [`lark-doc-update-workflow.md`](references/style/lark-doc-update-workflow.md)
+3. **创建或编辑文档内容** → 必读 [`lark-doc-xml.md`](references/lark-doc-xml.md)（XML 语法规则，仅当用户明确要求 Markdown 时改读 [`lark-doc-md.md`](references/lark-doc-md.md)）；从零创建时加读 [`lark-doc-create-workflow.md`](references/style/lark-doc-create-workflow.md)；编辑已有文档时加读 [`lark-doc-update.md`](references/lark-doc-update.md) 和 [`lark-doc-update-workflow.md`](references/style/lark-doc-update-workflow.md)
 4. **需要使用 callout、grid、table、whiteboard 等富 block 时** → 参考 [`lark-doc-style.md`](references/style/lark-doc-style.md) 的元素能力说明。该文件不是固定模板或强制排版规范；除非用户明确要求美化、重排版或特定风格，不要为了“达标”主动套用固定结构。
 
 **未读完以上文件就执行相应操作会导致参数选择错误或格式错误。**
@@ -36,11 +36,9 @@ lark-cli docs +update --api-version v2 --doc "文档URL或token" --command appen
 > - **精准编辑场景**（`docs +update` 的 `str_replace` / `block_insert_after` / `block_replace` / `block_delete` / `block_move_after` 等局部精修指令）：优先使用 XML（`--doc-format xml`，即默认值）。XML 能稳定表达 block 结构和样式，局部精修更可控；不要因为 Markdown 更简单就自行切换。
 
 ## 快速决策
-- 用户需要“某个 block 的直达链接 / 锚点链接”时：返回 `文档基础 URL#block_id`。如果当前只有文档 URL 没有 block_id，先用 `docs +fetch --detail with-ids` 拿到目标 block 的 id
-- 例：
-  - 已知文档 URL = `https://xxx.feishu.cn/docx/doxcn123`
-  - 已知 block_id = `blkcn456`
-  - 应返回 `https://xxx.feishu.cn/docx/doxcn123#blkcn456`
+- 先判定任务路径：找文档 / 导入导出走 [`lark-drive`](../lark-drive/SKILL.md)；只读 / 摘要用 `docs +fetch` 默认 `simple`；明确旧文本 → 新文本直接 `str_replace`；只有 block 链接、评论锚点、插入 / 替换 / 删除 / 移动才局部 fetch `with-ids`；保真改写已有内容才读 `full`
+- block 直达链接格式：`文档基础 URL#block_id`；没有 block_id 时局部 fetch `with-ids`
+- 连续执行多个文档写操作时，必须按 [`lark-doc-update.md`](references/lark-doc-update.md) 的「Block ID 生命周期」判断旧 block ID 是否还能复用；`overwrite` / `block_replace` / `block_delete` 后不要复用受影响的旧 ID，插入 / 复制后要重新 fetch 才能拿到新 block ID
 - 用户需要在文档内**创建、复制或移动**资源块（画板、电子表格、多维表格等）时，必须先读取 [`lark-doc-xml.md`](references/lark-doc-xml.md) 的「三、资源块」章节
 - 写文档时，由内容和用户意图决定表达形式；流程、架构、路线图、关键指标等信息可以使用画板，但不要默认把重要信息都画板化
 - 新增画板必须隔离到 SubAgent：简单图由 SubAgent 直接插入 `<whiteboard type="svg">完整 SVG</whiteboard>`，不读 `lark-whiteboard`；复杂图才由主 Agent 先建 `<whiteboard type="blank"></whiteboard>`，再启动 SubAgent 读取 `lark-whiteboard` 写入

skills/lark-doc/references/lark-doc-create.md

@@ -4,7 +4,7 @@
 > 1. [`lark-doc-xml.md`](lark-doc-xml.md) — XML 语法规则（使用 Markdown 格式时改读 [`lark-doc-md.md`](lark-doc-md.md)）
 > 2. [`lark-doc-create-workflow.md`](style/lark-doc-create-workflow.md) — 从零创作工作流（Code-Act Loop、并行执行策略）
 >
-> **需要使用 callout、grid、table、whiteboard 等富 block，或用户明确要求美化/重排版时，再参考 [`lark-doc-style.md`](style/lark-doc-style.md)。该文件是表达组件参考，不是固定模板。**
+> **需要富 block 或用户明确要求美化/重排版时，再参考 [`lark-doc-style.md`](style/lark-doc-style.md)。**
 >
 > **未读完以上文件就生成内容会导致格式错误。**
 
@@ -74,7 +74,7 @@ lark-cli docs +create --api-version v2 --doc-format markdown --content $'# 项
 ## 最佳实践
 
 - 文档标题从内容中自动提取：XML 使用 `<title>`；Markdown 使用文档开头唯一的一级标题（`# 标题`），正文从 `##` 开始。不要在内容开头重复写标题，也不要在 Markdown 正文中使用多个一级标题。
-- **创建较长的文档时只建骨架**：`--content` 仅传标题 + 各级 heading + 简短占位摘要；正文留给后续 `block_insert_after --block-id <章节标题 block_id>` 分段追加。一次性塞超长 `--content` 既容易触发参数限制，调试也更难。
+- **较长文档**：参考 [`lark-doc-create-workflow.md`](style/lark-doc-create-workflow.md) 先建骨架再分段写入；短文档可一次写完整内容。
 - **表达形式**：由用户目标和内容决定。需要结构化表达时可参考 [`lark-doc-style.md`](style/lark-doc-style.md)，但不要默认套用固定开头、固定富 block 比例或固定图表
 
 ## 参考

skills/lark-doc/references/lark-doc-update.md

@@ -5,7 +5,7 @@
 > 1. [`lark-doc-xml.md`](lark-doc-xml.md) — XML 语法规则（使用 Markdown 格式时改读 [`lark-doc-md.md`](lark-doc-md.md)）
 > 2. [`lark-doc-update-workflow.md`](style/lark-doc-update-workflow.md) — 改写增强工作流（Code-Act Loop、并行执行策略）
 >
-> **需要使用 callout、grid、table、whiteboard 等富 block，或用户明确要求美化/重排版时，再参考 [`lark-doc-style.md`](style/lark-doc-style.md)。该文件是表达组件参考，不是固定模板。**
+> **需要富 block 或用户明确要求美化/重排版时，再参考 [`lark-doc-style.md`](style/lark-doc-style.md)。**
 >
 > **未读完以上文件就生成内容会导致格式错误。**
 
@@ -44,6 +44,15 @@
 | `append` | ⚠️ 在文档**末尾**追加内容（等价于 `block_insert_after --block-id -1`）。**不适用于逐章填充**——逐章写入请用 `block_insert_after` 并指定对应标题的 `--block-id` | `--content` |
 | `block_move_after` | 移动已有 block 到指定位置 | `--block-id` `--src-block-ids` |
 
+## Block ID 生命周期
+
+写操作后不要默认复用之前 fetch 到的 block ID：
+
+- `overwrite` / `block_replace` / `block_delete`：受影响旧 ID 失效，继续 block 级操作前重新 fetch
+- `block_insert_after` / `append` / `block_copy_insert_after`：锚点 / 源 ID 通常保留，新内容是新 ID；要操作新内容先重新 fetch
+- `block_move_after`：被移动 ID 通常保留，但位置、章节、range 语义变化；后续依赖位置时重新 fetch
+- `str_replace`：简单行内替换通常不改变 ID；跨行 / 大段替换后如继续 block 级操作，先重新 fetch
+
 ## 指令示例
 
 ### str_replace — 全文文本替换
@@ -114,8 +123,6 @@ lark-cli docs +update --api-version v2 --doc "<doc_id>" --command block_replace
   --content '<p>替换后的段落内容</p>'
 ```
 
-> `block_replace` 由服务端执行整块替换，目标 block 的 ID 不保证在替换后继续可用。后续如果还要在替换后的块附近继续 `block_insert_after`、`range` 或其他 block 级操作，先重新 `docs +fetch --detail with-ids` 获取最新 block ID，不要复用旧 ID。
-
 ### block_delete — 删除指定 block
 
 ```bash
@@ -237,7 +244,6 @@ lark-cli docs +update --api-version v2 --doc "<doc_id>" --command str_replace \
 - **保护不可重建的内容**：图片、画板、电子表格等以 token 形式存储，替换时避开这些 block
 - **str_replace 的 replacement 支持富文本**：可以用行内标签 `<b>`、`<a>`、`<cite>`、`<latex>` 等替换普通文本为富文本
 - **同一 block 只能被 replace 一次**：多次修改同一 block 请合并为一次 block_replace
-- **block_replace 后重新获取 ID**：`block_replace` 成功后旧 block ID 不保证继续可用；继续做相邻块操作前，重新 `docs +fetch --detail with-ids`
 - **block_delete 支持批量**：用逗号分隔多个 block_id 一次删除
 - **复杂结构重组**：将多个段落转换为 grid / table 等复杂布局时，分步操作比 overwrite 更安全：
   1. 用 `block_insert_after` 在目标位置插入新的富文本结构

skills/lark-doc/references/lark-doc-whiteboard.md

@@ -9,11 +9,11 @@
 | `lark-doc`        | 识别画板机会、使用 Mermaid/SVG 创建图表、调度 SubAgent、插入简单 SVG 画板或复杂空白画板 | 主 Agent 不直接创作画板内容；              |
 | `lark-whiteboard` | 查询/导出已有画板；复杂图表生成（Mermaid/DSL/SVG 路由、场景选型、渲染验证）；写入已有/空白画板  | 仅特别复杂的图表或已有画板更新时由独立 SubAgent 读取 |
 
-## 画板优先规则
+## 画板适用规则
 
-写文档时，重要信息优先画板化。遇到核心流程、系统架构、方案对比、风险链路、里程碑、指标趋势、因果归因、组织关系、能力分层等内容，不要只用段落或表格承载；除非内容只是一次性补充说明，否则应规划为画板。
+写文档时，核心流程、系统架构、方案对比、风险链路、里程碑、指标趋势、因果归因、组织关系、能力分层等内容，如果图示能明显降低理解成本，可以规划为画板；结构简单或文字更清楚的内容不必强行画板化。
 
-同一篇文档可以有多个画板。优先设计多个聚焦画板，而不是把所有信息塞进一张大图。
+同一篇文档可以有多个画板。确有多个独立图示点时，可拆成多个聚焦画板，而不是把所有信息塞进一张大图。
 
 ## 文档与画板协同流程
 

skills/lark-doc/references/style/lark-doc-create-workflow.md

@@ -43,7 +43,7 @@
 8. **优先处理步骤三识别出的画板需求**：
    参考 [lark-doc-whiteboard.md](../lark-doc-whiteboard.md)中的方式，插入图表画板。
 9. Spawn 内容改写 Agent 定向润色：
-   - 文字密集且不易读的章节可转为 `<table>`/`<grid>`/`<callout>`，也可以拆段、改列表或保留纯文本
+   - 文字密集且不易读时，优先拆段、改列表、增加小标题或调整顺序；只有确实存在行列数据、并列对比或强提醒信息时，才考虑 `<table>` / `<grid>` / `<callout>`
    - 需要明显分隔的主题可补充 `<hr/>`，不强制章节间都使用
    - 本地图片使用 `docs +media-insert` 插入
 

skills/lark-doc/references/style/lark-doc-style.md

@@ -10,18 +10,18 @@
 2. **尊重用户风格**：用户给出样例、语气、结构或已有文档时，优先沿用；没有要求时不强行使用固定开头、固定章节或固定视觉组件
 3. **适度结构化**：结构化 block 用于降低理解成本，不为了“丰富”而堆叠
 4. **保持一致但不过度统一**：同类信息可使用相近表达，但允许因内容差异采用不同形式
-5. **重要信息画板化**：核心流程、架构、对比、风险、路线图、指标趋势等重要信息优先使用画板表达
+5. **图示服务理解**：流程、架构、对比、风险、路线图、指标趋势等内容在图示明显降低理解成本时，可使用画板表达
 
 ## 二、元素选择指南
 
 需要图表时，按类型选择插入方式：思维导图/时序图/类图/饼图/甘特图可用 `<whiteboard type="mermaid">` 直接内嵌；其他新图表可启动 SubAgent 插入 `<whiteboard type="svg">完整 SVG</whiteboard>`；只有编辑**已有**画板时才调用 **lark-whiteboard** skill。
 
-| 场景                                         | 推荐方案                                  |
+| 场景                                         | 可选表达方式                                |
 |--------------------------------------------|---------------------------------------|
-| 需要突出的一小段结论 / 摘要 / 注意事项                    | `<callout>`；是否使用 emoji 和颜色由文档语气决定 |
-| 方案对比 / 优劣势 / Before vs After                | `<grid>` 2 列分栏、`<table>` 或画板，按复杂度选择 |
+| 少数需要视觉提醒的短句，如风险、限制、待确认事项或关键提醒          | 需要视觉提醒时可用 `<callout>`；普通结论、摘要或章节导语优先使用段落、列表、小标题或加粗 |
+| 方案对比 / 优劣势 / Before vs After                | 简短对比可用段落、列表或 `<grid>`；维度较多且需要逐项比较时再考虑 `<table>` 或画板 |
 | 简短低风险对比                                    | `<grid>` 2 列分栏                        |
-| 3+ 属性的结构化数据 / 指标表                          | `<table>` + 表头背景色                     |
+| 需要按行列精确比较或查阅的数据，如指标、清单、字段说明、排期           | 可用 `<table>`；短要点、步骤、摘要或普通说明优先使用段落、列表或小标题 |
 | 任务清单 / 检查项                                 | `<checkbox>`                          |
 | 代码片段                                       | `<pre lang="x" caption="说明">`         |
 | 引用 / 公式                                    | `<blockquote>` / `<latex>`            |

skills/lark-doc/references/style/lark-doc-update-workflow.md

@@ -34,7 +34,7 @@
    参考 [lark-doc-whiteboard.md](../lark-doc-whiteboard.md)中的方式，插入图表画板。
 6. Spawn 内容改写 Agent 在不重叠的章节上并行改进，各 Agent 收到文档 token 和特定 block ID：
    - 沿用或轻微调整已有文档风格，除非用户要求彻底重排版
-   - 可以通过重写段落、调整标题、拆分列表、补表格/分栏/callout 等方式提升可读性
+   - 优先通过重写段落、调整标题、拆分列表或补充小标题提升可读性
    - 富 block 是可选表达手段，不因固定比例而添加；画板类需求只走第 5 步
 
 ### 步骤三：验证（串行）

skills/lark-vc-agent/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: lark-vc-agent
 version: 1.0.0
-description: "飞书视频会议会中能力：用于让应用机器人真实加入或离开正在进行的会议，并读取当前身份可见的会中事件，如参会人加入/离开、发言、聊天、屏幕共享。适用于用户询问正在开的会议发生了什么、谁在发言、是否共享内容，或需要发现当前可读的进行中会议 ID。不负责已结束会议搜索、参会人快照、纪要、逐字稿或录制查询，这些使用 lark-vc 技能。"
+description: "飞书视频会议会中能力：用于让应用机器人真实加入或离开正在进行的会议，并读取当前身份可见的会中事件、发送会中文本消息或反馈表情。适用于用户询问正在开的会议发生了什么、谁在发言、是否共享内容，或需要发现当前可读的进行中会议 ID。不负责已结束会议搜索、参会人快照、纪要、逐字稿或录制查询，这些使用 lark-vc 技能。"
 metadata:
   requires:
     bins: ["lark-cli"]
@@ -26,7 +26,7 @@ metadata:
 本 skill 与 [`lark-vc`](../lark-vc/SKILL.md) 并列：
 
 - **`lark-vc`** **负责"会后查询"**：搜索历史会议、参会人快照、纪要/逐字稿/录制
-- **`lark-vc-agent`** **负责"会中动作"**：机器人入会 / 读取进行中会议的实时事件 / 机器人离会
+- **`lark-vc-agent`** **负责"会中动作"**：机器人入会 / 读取进行中会议的实时事件 / 发送会中文本或反馈 / 机器人离会
 
 按此分工路由，避免两个 skill 语义混淆。
 
@@ -35,6 +35,7 @@ metadata:
 | "帮我入会 123456789"、"代我参会"、"让机器人进会旁听"                         | **本 skill** `+meeting-join`                                                                                                                           |
 | "会议现在还开着，谁刚加入了"、"会议里谁在发言"、"有人共享屏幕吗"（**进行中会议**）             | **本 skill** `+meeting-events`                                                                                                                         |
 | "我/某个用户现在在哪个会里"、"给我找当前可拉事件的 meeting_id"                         | **本 skill** `+meeting-list-active`                                                                                                                     |
+| "在会里发一句 xx"、"提示大家 xx"、"反馈听不到/看不到/声音清楚/效果不错"（**进行中会议**） | **本 skill** `+meeting-message-send`                                                                                                                     |
 | "退出会议"、"让机器人离开"                                            | **本 skill** `+meeting-leave`                                                                                                                          |
 | "昨天那场会有谁参加过"、"搜昨天的会"、"查纪要/逐字稿/录制"                          | [`lark-vc`](../lark-vc/SKILL.md)                                                                                                                      |
 | "帮我参会，结束后把纪要发到群" 等跨阶段场景                                    | 按序编排：本 skill（入会 → 读事件）→ 会议结束后用 [`lark-vc`](../lark-vc/SKILL.md) / [`lark-minutes`](../lark-minutes/SKILL.md) 拉纪要 → [`lark-im`](../lark-im/SKILL.md) 发群 |
@@ -49,7 +50,7 @@ metadata:
 | 查询目标用户且应用机器人也在会中的会议 | `--as bot --user-id <user_open_id>` | `--user-id` 必须是 `ou_...`；拿到的 `meeting_id` 后续继续用 `--as bot` 读事件 |
 | 用户明确要求应用机器人入会/旁听/代参会 | `--as bot` | 这是写操作，会真实产生入会记录；返回的 `meeting.id` 后续继续用 `--as bot` |
 
-硬规则：`meeting_id` 从哪种身份路径拿到，后续 `+meeting-events` 就沿用哪种身份，除非用户明确要求切换场景（例如从“仅查询我当前会”改成“让应用机器人入会旁听”）。
+硬规则：`meeting_id` 从哪种身份路径拿到，后续 `+meeting-events` / `+meeting-message-send` 就沿用哪种身份，除非用户明确要求切换场景（例如从“仅查询我当前会”改成“让应用机器人入会旁听”）。
 
 ## 核心场景
 
@@ -79,14 +80,31 @@ metadata:
 10. 用户直接问“这个会议讲了什么 / 现在讲到哪了”且上下文没有明确 `meeting_id` 时，先用用户身份发现当前会议；如果用户明确要求应用机器人视角，或上下文已经是应用机器人参会流程，再用应用身份发现。若返回多个会议，展示候选并让用户选择。
 11. 用户直接提供 **9 位会议号** 并询问会中事件/会议内容时，默认把它当作 active meeting 的筛选条件：先按当前身份查 active meetings，并在返回里匹配 `meeting_no == <9位会议号>`；匹配到唯一会议后取长数字 `meeting_id`，再用同一身份查事件。只有用户明确要求“入会 / 让应用机器人旁听 / 代我参会”时才改用 `+meeting-join`。
 
-### 3. 离开会议（写操作）
+### 3. 发送会中文本或反馈表情（写操作）
+
+1. 用户明确要求在当前进行中的会议里发送提示、说明、反馈或表情时，用 `+meeting-message-send`。
+2. 输入是长数字 `meeting_id`，不是 9 位会议号。若用户只给 9 位会议号，先按当前身份执行 `+meeting-list-active` 并按 `meeting_no` 匹配，匹配到唯一会议后再发送；不要为了发消息自动入会。
+3. 身份必须延续：`meeting_id` 来自用户身份发现，就继续 `--as user`；来自应用身份发现或应用机器人入会，就继续 `--as bot`。
+4. 文本消息使用 `--text`；反馈表情使用 `--emoji-type`。不要把自然语言硬编码转换成 key，但可以基于用户语义选择最匹配的已支持 key。
+5. 当前支持的会中特定反馈 key：`VC_CanNotSee`、`VC_NoSound`、`VC_LooksGood`、`VC_SoundsClear`。
+6. 该命令只暴露会中文本和会中反馈表情，不作为“发送绑定群消息”的默认能力；如果用户明确要发群聊，请路由到 [`lark-im`](../lark-im/SKILL.md)。
+7. 若使用应用身份发送，应用机器人必须在会中；若使用用户身份发送，当前用户必须正在该会议中。权限错误时按“应用身份权限配置检查”或“用户身份被拒绝时”处理。
+
+示例：
+
+```bash
+lark-cli vc +meeting-message-send --as user --meeting-id <meeting_id> --text "稍等，我在看文档"
+lark-cli vc +meeting-message-send --as bot --meeting-id <meeting_id> --msg-type reaction --emoji-type VC_NoSound
+```
+
+### 4. 离开会议（写操作）
 
 1. 只有用户明确要求机器人退出 / 离开 / 结束参会时，才用应用身份执行 `+meeting-leave --as bot --meeting-id <长数字 meeting_id>`；不应因任务完成而执行离会。
 2. `--meeting-id` **必须**是长数字会议 ID，通常来自 `+meeting-join` 返回的 `meeting.id`，也可以来自应用身份 `+meeting-list-active` 返回的 `meeting_id`。如果来自 list-active，必须确认应用机器人当前就在该会中。**不接受 9 位会议号**。
 3. 离会**立即生效**，机器人从会议的参会人列表中消失，对其他参会人可见；若需要重新入会，再跑一次 `+meeting-join` 即可（非真正"不可逆"）。
 4. 使用与入会或 active meeting 发现相同的应用身份离会。
 
-### 4. 获取当前可用的进行中会议 ID（读操作）
+### 5. 获取当前可用的进行中会议 ID（读操作）
 
 1. `+meeting-list-active` 用来发现当前进行中的会议，并拿到后续 `+meeting-events` 需要的长数字 `meeting_id`。
 2. 用户身份：`lark-cli vc +meeting-list-active --as user --format json`，用于发现当前登录用户正在参加的会议；后续 `+meeting-events` 继续 `--as user`。
@@ -95,7 +113,7 @@ metadata:
 5. 如果返回多个会议，不要自动任选一个；按 `meeting_title` / `meeting_no` / `meeting_id` 展示候选，等待用户明确选择后再调用 `+meeting-events`。
 6. 如果用户给了 9 位会议号，先在 active meeting 结果中按 `meeting_no` 匹配。匹配失败时，不要自动入会；只有用户明确要求应用机器人真实入会时，才询问或执行 `+meeting-join`。
 
-### 5. Agent 参会示范
+### 6. Agent 参会示范
 
 ```bash
 # 1. 入会，捕获 meeting.id
@@ -136,11 +154,13 @@ Shortcut 是对常用操作的高级封装（`lark-cli vc +<verb> [flags]`）。
 | [`+meeting-join`](references/lark-vc-agent-meeting-join.md)     | 写  | Join an in-progress meeting by 9-digit meeting number                      |
 | [`+meeting-list-active`](references/lark-vc-agent-meeting-list-active.md) | 读  | List active meetings and discover meeting_id for event reads               |
 | [`+meeting-events`](references/lark-vc-agent-meeting-events.md) | 读  | List meeting events visible to the app agent (participant joined/left, transcript, chat, share) |
+| [`+meeting-message-send`](references/lark-vc-agent-meeting-message-send.md) | 写  | Send an in-meeting text message or feedback reaction                       |
 | [`+meeting-leave`](references/lark-vc-agent-meeting-leave.md)   | 写  | Leave a meeting by meeting\_id                                             |
 
 - [`+meeting-join`](references/lark-vc-agent-meeting-join.md)：入参格式、写操作可见性风险、入会失败排查。
 - [`+meeting-list-active`](references/lark-vc-agent-meeting-list-active.md)：用户身份和应用身份的不同返回范围。
 - [`+meeting-events`](references/lark-vc-agent-meeting-events.md)：`meeting_id` 来源、身份延续、分页和错误码（10005 / 20001 / 20002）。
+- [`+meeting-message-send`](references/lark-vc-agent-meeting-message-send.md)：会中文本、会中反馈表情、身份延续和写操作风险。
 - [`+meeting-leave`](references/lark-vc-agent-meeting-leave.md)：`meeting_id` 的来源与写操作可见性。
 
 ## 应用身份权限配置检查

skills/lark-vc-agent/references/lark-vc-agent-meeting-message-send.md

@@ -0,0 +1,84 @@
+# vc +meeting-message-send
+
+发送会中文本消息或会中特定反馈表情。
+
+本 skill 对应 shortcut：`lark-cli vc +meeting-message-send`（调用 `POST /open-apis/vc/v1/meetings/{meeting_id}/messages`）。
+
+## 适用场景
+
+- 用户要求“在会里发一句话”“提示大家”“给当前会议发消息”。
+- 用户要求表达会中反馈，例如“听不到”“看不到”“声音清楚”“效果不错”。
+- 只用于正在进行中的会议；已结束会议不支持。
+
+## 身份规则
+
+`meeting_id` 从哪种身份路径拿到，发送消息时就沿用哪种身份：
+
+| meeting_id 来源 | 发送时身份 |
+| --- | --- |
+| `+meeting-list-active --as user` | `+meeting-message-send --as user` |
+| `+meeting-list-active --as bot --user-id <user_open_id>` | `+meeting-message-send --as bot` |
+| `+meeting-join --as bot` 返回的 `meeting.id` | `+meeting-message-send --as bot` |
+
+不要把用户身份发现的 `meeting_id` 改用应用身份发送，也不要把应用身份发现的 `meeting_id` 改用用户身份发送，除非用户明确要求切换。
+
+## 参数
+
+| 参数 | 说明 |
+| --- | --- |
+| `--meeting-id` | 必填，长数字 `meeting_id`，不是 9 位会议号 |
+| `--msg-type` | 可选，`text` 或 `reaction`；只传 `--text` 或只传 `--emoji-type` 时可自动推断 |
+| `--text` | 文本消息内容 |
+| `--emoji-type` | 会中反馈表情 key |
+| `--uuid` | 可选，幂等 key；不传则服务端生成 |
+
+## 文本消息
+
+```bash
+lark-cli vc +meeting-message-send --as user --meeting-id <meeting_id> --text "稍等，我在看文档"
+```
+
+文本消息会出现在会议内的文本互动区。不要把它当成绑定群消息发送能力；如果用户明确要求发到群聊，路由到 `lark-im`。
+
+## 反馈表情
+
+当前支持的会中特定反馈 key：
+
+| 用户表达 | 推荐 key |
+| --- | --- |
+| 听不到、没声音 | `VC_NoSound` |
+| 看不到、画面有问题 | `VC_CanNotSee` |
+| 声音清楚 | `VC_SoundsClear` |
+| 效果不错、看起来可以 | `VC_LooksGood` |
+
+```bash
+lark-cli vc +meeting-message-send --as bot --meeting-id <meeting_id> --msg-type reaction --emoji-type VC_NoSound
+```
+
+不要维护自己的自然语言到 key 的硬编码表；根据用户当前语义选择最匹配的 key。如果没有匹配项，先向用户确认，不要发送普通 IM 表情 key。
+
+## 9 位会议号处理
+
+如果用户给的是 9 位会议号并要求发送会中消息：
+
+1. 先按当前身份执行 `+meeting-list-active`。
+2. 在返回结果中按 `meeting_no` 匹配该 9 位会议号。
+3. 匹配到唯一会议后取长数字 `meeting_id`。
+4. 用发现该会议时的同一身份执行 `+meeting-message-send`。
+
+匹配失败时不要自动入会。只有用户明确要求“让应用机器人入会/旁听/代参会”时，才改用 `+meeting-join`。
+
+## 权限和前置条件
+
+- 用户身份：当前用户必须正在该会议中。
+- 应用身份：应用机器人必须正在该会议中。
+- 会议需要开启会中智能体/Agent 能力开关。
+- 需要 `vc:meeting.message:write` 权限；应用身份还需要应用已安装、数据范围已配置。
+
+应用身份权限错误时，不要引导用户反复 `auth login`。按主 skill 的“应用身份权限配置检查”处理。
+
+## 相关
+
+- [lark-vc-agent-meeting-list-active](lark-vc-agent-meeting-list-active.md) — 发现当前进行中会议 ID
+- [lark-vc-agent-meeting-events](lark-vc-agent-meeting-events.md) — 读取会中事件
+- [lark-vc-agent-meeting-join](lark-vc-agent-meeting-join.md) — 应用机器人入会