diff --git a/extensions/discord/skills/discord/SKILL.md b/extensions/discord/skills/discord/SKILL.md
index 10186a04d993..7e738d44b5fd 100644
--- a/extensions/discord/skills/discord/SKILL.md
+++ b/extensions/discord/skills/discord/SKILL.md
@@ -1,136 +1,43 @@
 ---
 name: discord
-description: "Discord message-tool ops: send/read/edit/delete, react, poll, pin, thread, search, presence, media/components."
+description: "Discord messaging workflows through OpenClaw's message tool."
 metadata: { "openclaw": { "emoji": "🎮", "requires": { "config": ["channels.discord.token"] } } }
 allowed-tools: ["message"]
 ---
 
 # Discord
 
-Use the `message` tool with `channel: "discord"`. No separate Discord tool.
+Use the `message` tool with `channel: "discord"`. The tool schema lists the actions enabled by the current account's `channels.discord.actions.*` gates; do not assume unavailable actions.
 
-## Rules
+## Workflow
 
-- Respect `channels.discord.actions.*` gates.
-- Prefer explicit `guildId`, `channelId`, `messageId`, `userId`.
-- Multi-account: pass `accountId` when needed.
-- Send targets: `to: "channel:<id>"` or `to: "user:<id>"`.
-- Mention users as `<@USER_ID>`.
-- Avoid Markdown tables in outbound Discord messages.
-- Prefer components v2 for rich UI; do not mix v2 `components` with legacy `embeds`.
+- Prefer stable `guildId`, `channelId`, `messageId`, and `userId` values from context. Pass `accountId` when more than one Discord account could apply.
+- Resolve the exact message before editing, deleting, pinning, moderating, or reacting when the user's reference is ambiguous.
+- Keep thread replies in their existing thread. A forum parent cannot receive components; send components to the created forum thread instead.
+- Confirm destructive or moderation actions unless the user already specified the exact target and action.
 
-## Common actions
+## Interactive components
 
-Send:
-
-```json
-{ "action": "send", "channel": "discord", "to": "channel:123", "message": "hello", "silent": true }
-```
-
-Send media:
+`components` must be a structured object or native component array, never a placeholder string. Do not combine components v2 with legacy `embeds`.
 
 ```json
 {
   "action": "send",
   "channel": "discord",
   "to": "channel:123",
-  "message": "see attachment",
-  "media": "file:///tmp/example.png"
+  "message": "Choose an option",
+  "components": {
+    "blocks": [
+      {
+        "type": "actions",
+        "buttons": [
+          { "label": "Approve", "style": "success", "callbackData": "approve" },
+          { "label": "Decline", "style": "danger", "callbackData": "decline" }
+        ]
+      }
+    ]
+  }
 }
 ```
 
-Components v2:
-
-```json
-{
-  "action": "send",
-  "channel": "discord",
-  "to": "channel:123",
-  "message": "Status",
-  "components": "[Carbon v2 components]"
-}
-```
-
-React:
-
-```json
-{ "action": "react", "channel": "discord", "channelId": "123", "messageId": "456", "emoji": "👍" }
-```
-
-Read:
-
-```json
-{ "action": "read", "channel": "discord", "to": "channel:123", "limit": 20 }
-```
-
-Edit/delete:
-
-```json
-{
-  "action": "edit",
-  "channel": "discord",
-  "channelId": "123",
-  "messageId": "456",
-  "message": "fixed typo"
-}
-```
-
-```json
-{ "action": "delete", "channel": "discord", "channelId": "123", "messageId": "456" }
-```
-
-Poll:
-
-```json
-{
-  "action": "poll",
-  "channel": "discord",
-  "to": "channel:123",
-  "pollQuestion": "Lunch?",
-  "pollOption": ["Pizza", "Sushi"],
-  "pollDurationHours": 24
-}
-```
-
-Pin:
-
-```json
-{ "action": "pin", "channel": "discord", "channelId": "123", "messageId": "456" }
-```
-
-Thread:
-
-```json
-{
-  "action": "thread-create",
-  "channel": "discord",
-  "channelId": "123",
-  "messageId": "456",
-  "threadName": "bug triage"
-}
-```
-
-Search:
-
-```json
-{
-  "action": "search",
-  "channel": "discord",
-  "guildId": "999",
-  "query": "release notes",
-  "channelIds": ["123"],
-  "limit": 10
-}
-```
-
-Presence, often gated:
-
-```json
-{
-  "action": "set-presence",
-  "channel": "discord",
-  "activityType": "playing",
-  "activityName": "OpenClaw",
-  "status": "online"
-}
-```
+Discord mention syntax, component availability, and form hints are injected automatically. Follow the current hints and tool schema rather than a duplicated action catalog.
diff --git a/extensions/feishu/skills/feishu-doc/SKILL.md b/extensions/feishu/skills/feishu-doc/SKILL.md
index db80fa8622a6..fab24d79699c 100644
--- a/extensions/feishu/skills/feishu-doc/SKILL.md
+++ b/extensions/feishu/skills/feishu-doc/SKILL.md
@@ -1,215 +1,40 @@
 ---
 name: feishu-doc
 description: |
-  Feishu document read/write operations. Activate when user mentions Feishu docs, cloud docs, or docx links.
+  Feishu document read/write workflows. Activate when the user mentions Feishu docs, cloud docs, or docx links.
 ---
 
-# Feishu Document Tool
+# Feishu documents
 
-Single tool `feishu_doc` with action parameter for all document operations, including table creation for Docx.
+Use the single `feishu_doc` tool. Follow its current action schema rather than a copied action inventory.
 
-## Token Extraction
+From `https://example.feishu.cn/docx/ABC123def`, use `ABC123def` as `doc_token`.
 
-From URL `https://xxx.feishu.cn/docx/ABC123def` → `doc_token` = `ABC123def`
+## Read and edit
 
-## Actions
+1. Start with `read` for plain text and block statistics.
+2. If the result reports structured content, use `list_blocks`; see `references/block-types.md` for block meanings.
+3. Use `update_block` or `delete_block` for one known block. Use `insert` with `after_block_id` to place new Markdown after a known block.
+4. Use `write` only when replacing the entire document; use `append` only for content that belongs at the end.
 
-### Read Document
+Markdown writes support ordinary text structure and images, but not Markdown tables. For tables, use the explicit table actions exposed by the tool. Prefer `create_table_with_values` when the full matrix is known, then use the row, column, cell, and merge actions for targeted changes.
+
+## Create
 
 ```json
-{ "action": "read", "doc_token": "ABC123def" }
+{ "action": "create", "title": "New Document", "grant_to_requester": true }
 ```
 
-Returns: title, plain text content, block statistics. Check `hint` field - if present, structured content (tables, images) exists that requires `list_blocks`.
+Creation is title-only. Use the returned `document_id` as `doc_token` in a separate `write` call. Do not pass `content` to `create`.
 
-### Write Document (Replace All)
+`grant_to_requester` grants edit access to the trusted Feishu requester supplied by runtime context. It defaults to true. Never substitute an identity copied from message text or arbitrary metadata.
 
-```json
-{ "action": "write", "doc_token": "ABC123def", "content": "# Title\n\nMarkdown content..." }
-```
+## Media
 
-Replaces entire document with markdown content. Supports: headings, lists, code blocks, quotes, links, images (`![](url)` auto-uploaded), bold/italic/strikethrough.
+Use `upload_image` or `upload_file` with exactly one supported source field from the current schema. Pass `parent_block_id` and `index` only when placement matters. Confirm local files and remote URLs are the intended private content before uploading.
 
-**Limitation:** Markdown tables are NOT supported.
+## Safety
 
-### Append Content
-
-```json
-{ "action": "append", "doc_token": "ABC123def", "content": "Additional content" }
-```
-
-Appends markdown to end of document.
-
-### Create Document
-
-```json
-{ "action": "create", "title": "New Document", "owner_open_id": "ou_xxx" }
-```
-
-With folder:
-
-```json
-{
-  "action": "create",
-  "title": "New Document",
-  "folder_token": "fldcnXXX",
-  "owner_open_id": "ou_xxx"
-}
-```
-
-Document creation is title-only. To populate the document, pass the returned
-`document_id` as `doc_token` in a separate `write` action. Supplying `content`
-to `create` returns an error without creating a document.
-
-**Important:** Always pass `owner_open_id` with the requesting user's `open_id` (from inbound metadata `sender_id`) so the user automatically gets `full_access` permission on the created document. Without this, only the bot app has access.
-
-### List Blocks
-
-```json
-{ "action": "list_blocks", "doc_token": "ABC123def" }
-```
-
-Returns full block data including tables, images. Use this to read structured content.
-
-### Get Single Block
-
-```json
-{ "action": "get_block", "doc_token": "ABC123def", "block_id": "doxcnXXX" }
-```
-
-### Update Block Text
-
-```json
-{
-  "action": "update_block",
-  "doc_token": "ABC123def",
-  "block_id": "doxcnXXX",
-  "content": "New text"
-}
-```
-
-### Delete Block
-
-```json
-{ "action": "delete_block", "doc_token": "ABC123def", "block_id": "doxcnXXX" }
-```
-
-### Create Table (Docx Table Block)
-
-```json
-{
-  "action": "create_table",
-  "doc_token": "ABC123def",
-  "row_size": 2,
-  "column_size": 2,
-  "column_width": [200, 200]
-}
-```
-
-Optional: `parent_block_id` to insert under a specific block.
-
-### Write Table Cells
-
-```json
-{
-  "action": "write_table_cells",
-  "doc_token": "ABC123def",
-  "table_block_id": "doxcnTABLE",
-  "values": [
-    ["A1", "B1"],
-    ["A2", "B2"]
-  ]
-}
-```
-
-### Create Table With Values (One-step)
-
-```json
-{
-  "action": "create_table_with_values",
-  "doc_token": "ABC123def",
-  "row_size": 2,
-  "column_size": 2,
-  "column_width": [200, 200],
-  "values": [
-    ["A1", "B1"],
-    ["A2", "B2"]
-  ]
-}
-```
-
-Optional: `parent_block_id` to insert under a specific block.
-
-### Upload Image to Docx (from URL or local file)
-
-```json
-{
-  "action": "upload_image",
-  "doc_token": "ABC123def",
-  "url": "https://example.com/image.png"
-}
-```
-
-Or local path with position control:
-
-```json
-{
-  "action": "upload_image",
-  "doc_token": "ABC123def",
-  "file_path": "/tmp/image.png",
-  "parent_block_id": "doxcnParent",
-  "index": 5
-}
-```
-
-Optional `index` (0-based) inserts the image at a specific position among sibling blocks. Omit to append at end.
-
-**Note:** Image display size is determined by the uploaded image's pixel dimensions. For small images (e.g. 480x270 GIFs), scale to 800px+ width before uploading to ensure proper display.
-
-### Upload File Attachment to Docx (from URL or local file)
-
-```json
-{
-  "action": "upload_file",
-  "doc_token": "ABC123def",
-  "url": "https://example.com/report.pdf"
-}
-```
-
-Or local path:
-
-```json
-{
-  "action": "upload_file",
-  "doc_token": "ABC123def",
-  "file_path": "/tmp/report.pdf",
-  "filename": "Q1-report.pdf"
-}
-```
-
-Rules:
-
-- exactly one of `url` / `file_path`
-- optional `filename` override
-- optional `parent_block_id`
-
-## Reading Workflow
-
-1. Start with `action: "read"` - get plain text + statistics
-2. Check `block_types` in response for Table, Image, Code, etc.
-3. If structured content exists, use `action: "list_blocks"` for full data
-
-## Configuration
-
-```yaml
-channels:
-  feishu:
-    tools:
-      doc: true # default: true
-```
-
-**Note:** `feishu_wiki` depends on this tool - wiki page content is read/written via `feishu_doc`.
-
-## Permissions
-
-Required: `docx:document`, `docx:document:readonly`, `docx:document.block:convert`, `drive:drive`
+- Resolve exact document and block IDs before destructive edits.
+- Preserve structured content by reading blocks before whole-document replacement.
+- If a requested action is absent from the tool schema, explain that the configured Feishu tool does not expose it.
diff --git a/extensions/feishu/skills/feishu-doc/references/block-types.md b/extensions/feishu/skills/feishu-doc/references/block-types.md
index 8ce599fe869a..97b9d8a17d84 100644
--- a/extensions/feishu/skills/feishu-doc/references/block-types.md
+++ b/extensions/feishu/skills/feishu-doc/references/block-types.md
@@ -1,6 +1,6 @@
 # Feishu Block Types Reference
 
-Complete reference for Feishu document block types. Use with `feishu_doc_list_blocks`, `feishu_doc_update_block`, and `feishu_doc_delete_block`.
+Complete reference for Feishu document block types. Use the single `feishu_doc` tool with its `list_blocks`, `update_block`, and `delete_block` actions.
 
 ## Block Type Table
 
@@ -59,10 +59,11 @@ Complete reference for Feishu document block types. Use with `feishu_doc_list_bl
 
 ### Text-based blocks (2-17, 19)
 
-Update text content using `feishu_doc_update_block`:
+Update text content with `feishu_doc`:
 
 ```json
 {
+  "action": "update_block",
   "doc_token": "ABC123",
   "block_id": "block_xxx",
   "content": "New text content"
@@ -71,13 +72,11 @@ Update text content using `feishu_doc_update_block`:
 
 ### Image blocks (27)
 
-Images cannot be updated directly via `update_block`. Use `feishu_doc_write` or `feishu_doc_append` with markdown to add new images.
+Images cannot be updated directly via `update_block`. Use `upload_image` to add a replacement image, then delete the old block only after the new upload succeeds.
 
 ### Table blocks (31)
 
-**Important:** Table blocks CANNOT be created via the `documentBlockChildren.create` API (error 1770029). This affects `feishu_doc_write` and `feishu_doc_append` - markdown tables will be skipped with a warning.
-
-Tables can only be read (via `list_blocks`) and individual cells (type 32) can be updated, but new tables cannot be inserted programmatically via markdown.
+Markdown tables are not converted by `write`, `append`, or `insert`. Create tables with `create_table` or `create_table_with_values`; use the table row, column, cell, and merge actions exposed by the current schema for later edits.
 
 ### Container blocks (24, 25, 35)
 
@@ -87,17 +86,13 @@ Grid and QuoteContainer are layout containers. Edit their child blocks instead.
 
 ### Replace specific paragraph
 
-1. `feishu_doc_list_blocks` - find the block_id
-2. `feishu_doc_update_block` - update its content
+1. Call `feishu_doc` with `action: "list_blocks"` to find the block ID.
+2. Call `feishu_doc` with `action: "update_block"` to replace its text.
 
 ### Insert content at specific location
 
-Currently, the API only supports appending to document end. For insertion at specific positions, consider:
-
-1. Read existing content
-2. Delete affected blocks
-3. Rewrite with new content in desired order
+Call `feishu_doc` with `action: "insert"`, the target `doc_token`, Markdown `content`, and the preceding block's ID as `after_block_id`.
 
 ### Delete multiple blocks
 
-Blocks must be deleted one at a time. Delete child blocks before parent containers.
+Call `delete_block` for exact block IDs. Delete child blocks before parent containers and confirm broad deletions first.
diff --git a/extensions/feishu/skills/feishu-drive/SKILL.md b/extensions/feishu/skills/feishu-drive/SKILL.md
index d8904f71a499..a93334376810 100644
--- a/extensions/feishu/skills/feishu-drive/SKILL.md
+++ b/extensions/feishu/skills/feishu-drive/SKILL.md
@@ -1,114 +1,28 @@
 ---
 name: feishu-drive
 description: |
-  Feishu cloud storage file management. Activate when user mentions cloud space, folders, drive.
+  Feishu cloud-storage and comment workflows. Activate when the user mentions cloud space, folders, Drive files, or document comments.
 ---
 
-# Feishu Drive Tool
+# Feishu Drive
 
-Single tool `feishu_drive` for cloud storage operations.
+Use the single `feishu_drive` tool and its current action schema.
 
-## Token Extraction
+From `https://example.feishu.cn/drive/folder/ABC123`, use `ABC123` as `folder_token`.
 
-From URL `https://xxx.feishu.cn/drive/folder/ABC123` → `folder_token` = `ABC123`
+## Files and folders
 
-## Actions
+- Start from a folder shared with the bot. Bot credentials normally have no usable personal root folder.
+- For paginated folder listings, keep the same `folder_token` and pass the returned `page_token` until no continuation token remains.
+- Use `info` with the exact file token and type returned by Drive or wiki discovery.
+- Resolve the exact source and destination before moving or deleting. Confirm destructive deletes when the user's intent or target is unclear.
+- Create subfolders inside a shared folder; creating at the account root normally fails for bots.
 
-### List Folder Contents
+## Comments
 
-```json
-{ "action": "list" }
-```
+- Use `list_comments`, then `list_comment_replies` with the exact `comment_id` to inspect a discussion.
+- Use `add_comment` for a document-level comment. Include `block_id` only when the user wants a comment on one known Docx block.
+- Use `reply_comment` for an existing comment thread.
+- Preserve the file type and pagination fields returned by the tool. The schema is authoritative for which file types each comment action accepts.
 
-Requests the account root (no `folder_token`). Bot credentials normally have no root folder, so
-use a folder that has been shared with the bot instead.
-
-```json
-{ "action": "list", "folder_token": "fldcnXXX", "page_size": 100 }
-```
-
-Returns one page of files with token, name, type, url, timestamps, and `next_page_token` when
-another page is available. To continue, pass the returned token with the same folder token:
-
-```json
-{
-  "action": "list",
-  "folder_token": "fldcnXXX",
-  "page_size": 100,
-  "page_token": "next-page-token"
-}
-```
-
-`page_size` must be between 1 and 200. Pagination requires a concrete shared `folder_token`;
-root-list cursors are not forwarded.
-
-### Get File Info
-
-```json
-{ "action": "info", "file_token": "ABC123", "type": "docx" }
-```
-
-Looks up file metadata directly by token and type, regardless of which shared folder contains it.
-Shortcuts are the exception: Feishu's metadata API does not support the `shortcut` type, so shortcut
-info retains the root-directory lookup behavior.
-
-`type`: `doc`, `docx`, `sheet`, `bitable`, `folder`, `file`, `mindnote`, `shortcut`
-
-### Create Folder
-
-```json
-{ "action": "create_folder", "name": "New Folder" }
-```
-
-In parent folder:
-
-```json
-{ "action": "create_folder", "name": "New Folder", "folder_token": "fldcnXXX" }
-```
-
-### Move File
-
-```json
-{ "action": "move", "file_token": "ABC123", "type": "docx", "folder_token": "fldcnXXX" }
-```
-
-### Delete File
-
-```json
-{ "action": "delete", "file_token": "ABC123", "type": "docx" }
-```
-
-## File Types
-
-| Type       | Description             |
-| ---------- | ----------------------- |
-| `doc`      | Old format document     |
-| `docx`     | New format document     |
-| `sheet`    | Spreadsheet             |
-| `bitable`  | Multi-dimensional table |
-| `folder`   | Folder                  |
-| `file`     | Uploaded file           |
-| `mindnote` | Mind map                |
-| `shortcut` | Shortcut                |
-
-## Configuration
-
-```yaml
-channels:
-  feishu:
-    tools:
-      drive: true # default: true
-```
-
-## Permissions
-
-- `drive:drive` - Full access (create, move, delete)
-- `drive:drive:readonly` - Read only (list and root-level info fallback)
-- `drive:drive.metadata:readonly` - Direct `info` lookup outside the root (not needed with `drive:drive`)
-
-## Known Limitations
-
-- **Bots have no root folder**: Feishu bots use `tenant_access_token` and don't have their own "My Space". The root folder concept only exists for user accounts. This means:
-  - `create_folder` without `folder_token` will fail (400 error)
-  - Bot can only access files/folders that have been **shared with it**
-  - **Workaround**: User must first create a folder manually and share it with the bot, then bot can create subfolders inside it
+Only expose or forward file and comment contents needed for the user's request; shared Drive data may be private.
diff --git a/extensions/feishu/skills/feishu-perm/SKILL.md b/extensions/feishu/skills/feishu-perm/SKILL.md
index 1ce5db8b86c5..b7e60caaf4ae 100644
--- a/extensions/feishu/skills/feishu-perm/SKILL.md
+++ b/extensions/feishu/skills/feishu-perm/SKILL.md
@@ -1,119 +1,19 @@
 ---
 name: feishu-perm
 description: |
-  Feishu permission management for documents and files. Activate when user mentions sharing, permissions, collaborators.
+  Feishu collaborator and permission workflows. Activate when the user explicitly asks to inspect or change sharing, permissions, or collaborators.
 ---
 
-# Feishu Permission Tool
+# Feishu permissions
 
-Single tool `feishu_perm` for managing file/document permissions.
+Use the single `feishu_perm` tool and its current action schema. This tool is disabled by default because it changes access to user data; if it is unavailable, explain that `channels.feishu.tools.perm` must be enabled.
 
-## Actions
+## Workflow
 
-### List Collaborators
+1. Resolve the exact file token and type.
+2. Use `list` to inspect current collaborators before changing access.
+3. For `add`, resolve the collaborator's exact identifier and choose the least permission that satisfies the request.
+4. For `remove`, confirm the exact collaborator and file when the request is ambiguous or broad.
+5. Report the resulting permission change without exposing unrelated collaborator data.
 
-```json
-{ "action": "list", "token": "ABC123", "type": "docx" }
-```
-
-Returns: members with member_type, member_id, perm, name.
-
-### Add Collaborator
-
-```json
-{
-  "action": "add",
-  "token": "ABC123",
-  "type": "docx",
-  "member_type": "email",
-  "member_id": "user@example.com",
-  "perm": "edit"
-}
-```
-
-### Remove Collaborator
-
-```json
-{
-  "action": "remove",
-  "token": "ABC123",
-  "type": "docx",
-  "member_type": "email",
-  "member_id": "user@example.com"
-}
-```
-
-## Token Types
-
-| Type       | Description             |
-| ---------- | ----------------------- |
-| `doc`      | Old format document     |
-| `docx`     | New format document     |
-| `sheet`    | Spreadsheet             |
-| `bitable`  | Multi-dimensional table |
-| `folder`   | Folder                  |
-| `file`     | Uploaded file           |
-| `wiki`     | Wiki node               |
-| `mindnote` | Mind map                |
-
-## Member Types
-
-| Type               | Description        |
-| ------------------ | ------------------ |
-| `email`            | Email address      |
-| `openid`           | User open_id       |
-| `userid`           | User user_id       |
-| `unionid`          | User union_id      |
-| `openchat`         | Group chat open_id |
-| `opendepartmentid` | Department open_id |
-
-## Permission Levels
-
-| Perm          | Description                          |
-| ------------- | ------------------------------------ |
-| `view`        | View only                            |
-| `edit`        | Can edit                             |
-| `full_access` | Full access (can manage permissions) |
-
-## Examples
-
-Share document with email:
-
-```json
-{
-  "action": "add",
-  "token": "doxcnXXX",
-  "type": "docx",
-  "member_type": "email",
-  "member_id": "alice@company.com",
-  "perm": "edit"
-}
-```
-
-Share folder with group:
-
-```json
-{
-  "action": "add",
-  "token": "fldcnXXX",
-  "type": "folder",
-  "member_type": "openchat",
-  "member_id": "oc_xxx",
-  "perm": "view"
-}
-```
-
-## Configuration
-
-```yaml
-channels:
-  feishu:
-    tools:
-      perm: true # default: false (disabled)
-```
-
-**Note:** This tool is disabled by default because permission management is a sensitive operation. Enable explicitly if needed.
-
-## Permissions
-
-Required: `drive:permission`
+Never infer an email, user ID, department, or chat from a display name alone. Follow the current schema for supported member types, token types, and permission levels.
diff --git a/extensions/feishu/skills/feishu-wiki/SKILL.md b/extensions/feishu/skills/feishu-wiki/SKILL.md
index ad68d8bffa1a..c87259d1ff5b 100644
--- a/extensions/feishu/skills/feishu-wiki/SKILL.md
+++ b/extensions/feishu/skills/feishu-wiki/SKILL.md
@@ -1,122 +1,26 @@
 ---
 name: feishu-wiki
 description: |
-  Feishu knowledge base navigation. Activate when user mentions knowledge base, wiki, or wiki links.
+  Feishu knowledge-base navigation workflows. Activate when the user mentions a knowledge base, wiki, or wiki link.
 ---
 
-# Feishu Wiki Tool
+# Feishu wiki
 
-Single tool `feishu_wiki` for knowledge base operations.
+Use the single `feishu_wiki` tool and its current action schema.
 
-Wiki `space_id` values are opaque strings. Always keep them quoted in tool calls, even when they contain only digits; passing a long numeric-looking ID as a number can corrupt the suffix due to JavaScript number precision limits.
+From `https://example.feishu.cn/wiki/ABC123def`, use `ABC123def` as `token`. Treat every `space_id` as an opaque quoted string, even when it contains only digits.
 
-## Token Extraction
+## Navigate
 
-From URL `https://xxx.feishu.cn/wiki/ABC123def` → `token` = `ABC123def`
+- Use `spaces` to enumerate accessible knowledge spaces and `nodes` for a space or parent node.
+- Continue pagination with the returned `page_token` while `has_more` is true, keeping the same space and parent.
+- Use `search` when the user provides a query but not an exact node.
+- Use `get` to resolve a wiki token to its `node_token`, `obj_token`, and `obj_type`.
 
-## Actions
+## Create and organize
 
-### List Knowledge Spaces
+Wiki creation supports only `docx`, `sheet`, and `bitable`; `docx` is the default. Resolve an exact space and parent before create, move, or rename operations, and confirm ambiguous or destructive reorganizations.
 
-```json
-{ "action": "spaces" }
-```
+## Wiki content workflow
 
-Returns one page of accessible wiki spaces plus `has_more` and `page_token`.
-Continue with the returned `page_token` while `has_more` is true:
-
-```json
-{ "action": "spaces", "page_token": "next-page-token" }
-```
-
-### List Nodes
-
-```json
-{ "action": "nodes", "space_id": "7xxx" }
-```
-
-With parent:
-
-```json
-{ "action": "nodes", "space_id": "7xxx", "parent_node_token": "wikcnXXX" }
-```
-
-Returns one page of nodes plus `has_more` and `page_token`. Continue with the
-same `space_id` and `parent_node_token`, adding the returned `page_token`, while
-`has_more` is true. Both list actions accept optional `page_size` from 1 to 50.
-
-### Get Node Details
-
-```json
-{ "action": "get", "token": "ABC123def" }
-```
-
-Returns: `node_token`, `obj_token`, `obj_type`, etc. Use `obj_token` with `feishu_doc` to read/write the document.
-
-### Create Node
-
-```json
-{ "action": "create", "space_id": "7xxx", "title": "New Page" }
-```
-
-With type and parent:
-
-```json
-{
-  "action": "create",
-  "space_id": "7xxx",
-  "title": "Sheet",
-  "obj_type": "sheet",
-  "parent_node_token": "wikcnXXX"
-}
-```
-
-`obj_type`: `docx` (default), `sheet`, `bitable`, `mindnote`, `file`, `doc`, `slides`
-
-### Move Node
-
-```json
-{ "action": "move", "space_id": "7xxx", "node_token": "wikcnXXX" }
-```
-
-To different location:
-
-```json
-{
-  "action": "move",
-  "space_id": "7xxx",
-  "node_token": "wikcnXXX",
-  "target_space_id": "7yyy",
-  "target_parent_token": "wikcnYYY"
-}
-```
-
-### Rename Node
-
-```json
-{ "action": "rename", "space_id": "7xxx", "node_token": "wikcnXXX", "title": "New Title" }
-```
-
-## Wiki-Doc Workflow
-
-To edit a wiki page:
-
-1. Get node: `{ "action": "get", "token": "wiki_token" }` → returns `obj_token`
-2. Read doc: `feishu_doc { "action": "read", "doc_token": "obj_token" }`
-3. Write doc: `feishu_doc { "action": "write", "doc_token": "obj_token", "content": "..." }`
-
-## Configuration
-
-```yaml
-channels:
-  feishu:
-    tools:
-      wiki: true # default: true
-      doc: true # required - wiki content uses feishu_doc
-```
-
-**Dependency:** This tool requires `feishu_doc` to be enabled. Wiki pages are documents - use `feishu_wiki` to navigate, then `feishu_doc` to read/edit content.
-
-## Permissions
-
-Required: `wiki:wiki` or `wiki:wiki:readonly`
+Wiki navigation is independent of the document tool. When `get` returns a `docx` object, use its `obj_token` as `doc_token` with `feishu_doc` to read or edit the page. Other object types require a currently available tool that supports that type; do not treat them as documents.
diff --git a/extensions/imessage/skills/imsg/SKILL.md b/extensions/imessage/skills/imsg/SKILL.md
index dc4b5ca72f5a..516f433746ae 100644
--- a/extensions/imessage/skills/imsg/SKILL.md
+++ b/extensions/imessage/skills/imsg/SKILL.md
@@ -25,7 +25,7 @@ metadata:
 
 # imsg
 
-Use `imsg` when an OpenClaw agent must act through the user's local macOS Messages.app account: inspect iMessage/SMS history, choose the correct DM or group, send messages/files, reply, react, vote in polls, or use private-API iMessage features.
+Use the generic `message` tool first for actions that its current iMessage schema exposes. Use `imsg` when the task needs local Messages.app history, target discovery, watching, or an administrative/private-API capability that is not exposed by `message`.
 
 Do not use this skill for Telegram, Signal, WhatsApp, Discord, Slack, or for replying inside the current OpenClaw conversation when the configured channel already routes the reply.
 
@@ -33,7 +33,7 @@ Do not use this skill for Telegram, Signal, WhatsApp, Discord, Slack, or for rep
 
 1. Resolve the conversation first.
 2. Choose DM, existing group, or new group.
-3. Pick the lowest-capability command that can do the requested action.
+3. Prefer an available `message` action; otherwise pick the lowest-capability `imsg` command that preserves the requested semantics.
 4. Confirm any send or visible state change unless the user already gave exact recipient, content, and action.
 5. Execute with stable identifiers: prefer `--chat-id` for normal sends/watch/history and `--chat` chat GUID for bridge actions.
 
@@ -74,7 +74,7 @@ Do not make `jq` a hard prerequisite for the skill; it is only a convenient form
 
 ## Capability Choice
 
-Use standard commands for reads, target resolution, and plain sends:
+Use `imsg` standard commands when the current `message` schema does not cover the operation or when local history is needed:
 
 - Read/list/search/watch: `chats`, `group`, `history`, `search`, `watch`
 - Basic text/file send: `send`
diff --git a/extensions/qqbot/skills/qqbot-remind/SKILL.md b/extensions/qqbot/skills/qqbot-remind/SKILL.md
index 6f8d07734bdd..f738140a5813 100644
--- a/extensions/qqbot/skills/qqbot-remind/SKILL.md
+++ b/extensions/qqbot/skills/qqbot-remind/SKILL.md
@@ -36,81 +36,7 @@ metadata: { "openclaw": { "emoji": "⏰", "requires": { "config": ["channels.qqb
 1. 调用 `qqbot_remind`：`{ "action": "add", "content": "喝水", "time": "5m" }`
 2. 工具返回成功后，回复用户：`⏰ 好的，5分钟后提醒你喝水~`
 
----
-
-## 备用方案（直接使用 `cron` 工具）
-
-> 仅当 `qqbot_remind` 工具不可用但 `cron` 工具可用时使用以下方式。
-
-### 核心规则
-
-> **payload.kind 必须是 `"agentTurn"`，绝对不能用 `"systemEvent"`！**
-> `systemEvent` 只在 AI 会话内部注入文本，用户收不到 QQ 消息。
-
-**不可更改字段**：
-
-| 字段                 | 固定值        | 原因                         |
-| -------------------- | ------------- | ---------------------------- |
-| `payload.kind`       | `"agentTurn"` | `systemEvent` 不会发 QQ 消息 |
-| `delivery.mode`      | `"announce"`  | 主动投递模式                 |
-| `delivery.channel`   | `"qqbot"`     | QQ 通道标识                  |
-| `delivery.to`        | 目标地址      | 从当前会话上下文获取         |
-| `delivery.accountId` | 当前账户 ID   | 多账号场景下不可省略         |
-| `sessionTarget`      | `"isolated"`  | 隔离会话避免污染             |
-
-> `schedule.at` 必须是**绝对 ISO 8601 时间戳**（如 `"2026-07-29T09:00:00.000Z"`），不支持 `"5m"` 等相对字符串。
-> 计算方式：`当前时间戳ms + 延迟毫秒`。
-
-### 一次性提醒（schedule.kind = "at"）
-
-```json
-{
-  "action": "add",
-  "job": {
-    "name": "{任务名}",
-    "schedule": { "kind": "at", "at": "{当前时间戳ms + N*60000 转换为 ISO 8601}" },
-    "sessionTarget": "isolated",
-    "wakeMode": "now",
-    "deleteAfterRun": true,
-    "payload": {
-      "kind": "agentTurn",
-      "message": "你是一个暖心的提醒助手。请用温暖、有趣的方式提醒用户：{提醒内容}。要求：(1) 不要回复HEARTBEAT_OK (2) 不要解释你是谁 (3) 直接输出一条暖心的提醒消息 (4) 可以加一句简短的鸡汤或关怀的话 (5) 控制在2-3句话以内 (6) 用emoji点缀"
-    },
-    "delivery": {
-      "mode": "announce",
-      "channel": "qqbot",
-      "to": "qqbot:c2c:{openid}",
-      "accountId": "{accountId}"
-    }
-  }
-}
-```
-
-### 周期提醒（schedule.kind = "cron"）
-
-```json
-{
-  "action": "add",
-  "job": {
-    "name": "{任务名}",
-    "schedule": { "kind": "cron", "expr": "0 8 * * *", "tz": "{用户确认的 IANA 时区}" },
-    "sessionTarget": "isolated",
-    "wakeMode": "now",
-    "payload": {
-      "kind": "agentTurn",
-      "message": "你是一个暖心的提醒助手。请用温暖、有趣的方式提醒用户：{提醒内容}。要求：(1) 不要回复HEARTBEAT_OK (2) 不要解释你是谁 (3) 直接输出一条暖心的提醒消息 (4) 可以加一句简短的鸡汤或关怀的话 (5) 控制在2-3句话以内 (6) 用emoji点缀"
-    },
-    "delivery": {
-      "mode": "announce",
-      "channel": "qqbot",
-      "to": "qqbot:c2c:{openid}",
-      "accountId": "{accountId}"
-    }
-  }
-}
-```
-
-> 周期任务**不加** `deleteAfterRun`。群聊 `delivery.to` 格式为 `"qqbot:group:{group_openid}"`。
+`qqbot_remind` 不可用时，不要绕过它直接创建 Gateway 任务。说明当前无法安全注册 QQ 提醒，并建议用户检查 QQBot 工具配置。
 
 ---
 
diff --git a/extensions/slack/skills/slack/SKILL.md b/extensions/slack/skills/slack/SKILL.md
index 711139bba54d..ac04a2336a64 100644
--- a/extensions/slack/skills/slack/SKILL.md
+++ b/extensions/slack/skills/slack/SKILL.md
@@ -1,78 +1,20 @@
 ---
 name: slack
-description: "Slack tool actions: send/read/edit/delete messages, react, pin/unpin, list pins/reactions/emoji, member info."
+description: "Slack messaging workflows through OpenClaw's message tool."
 metadata: { "openclaw": { "emoji": "💬", "requires": { "config": ["channels.slack"] } } }
+allowed-tools: ["message"]
 ---
 
 # Slack
 
-Use the `slack` tool. Reuse `channelId` and Slack timestamp message IDs from context when present.
+Use the `message` tool with `channel: "slack"`. The tool schema lists the actions enabled for the current Slack account; do not assume an action that is not present.
 
-## Inputs
+## Workflow
 
-- `channelId`: Slack channel ID.
-- `messageId`: Slack timestamp, e.g. `1712023032.1234`.
-- `to`: `channel:<id>` or `user:<id>` for sends.
-- `emoji`: Unicode or `:name:` for reactions.
+- Prefer stable Slack IDs from context. Sends outside the current conversation use `channel:<id>` or `user:<id>` targets.
+- Keep replies in the current thread unless the user asks for a top-level post. For another thread, pass its Slack timestamp as `threadId`; use the same timestamp as `messageId` for message-specific actions.
+- Before editing, deleting, pinning, or reacting to an ambiguous message, read the conversation and resolve its exact ID.
+- When multiple Slack accounts are configured, pass `accountId` rather than guessing.
+- Confirm destructive deletes when the target or intent is unclear.
 
-## Actions
-
-```json
-{ "action": "sendMessage", "to": "channel:C123", "content": "Hello" }
-```
-
-```json
-{ "action": "readMessages", "channelId": "C123", "limit": 20 }
-```
-
-```json
-{
-  "action": "react",
-  "channelId": "C123",
-  "messageId": "1712023032.1234",
-  "emoji": ":white_check_mark:"
-}
-```
-
-```json
-{ "action": "reactions", "channelId": "C123", "messageId": "1712023032.1234" }
-```
-
-```json
-{
-  "action": "editMessage",
-  "channelId": "C123",
-  "messageId": "1712023032.1234",
-  "content": "Updated text"
-}
-```
-
-```json
-{ "action": "deleteMessage", "channelId": "C123", "messageId": "1712023032.1234" }
-```
-
-```json
-{ "action": "pinMessage", "channelId": "C123", "messageId": "1712023032.1234" }
-```
-
-```json
-{ "action": "unpinMessage", "channelId": "C123", "messageId": "1712023032.1234" }
-```
-
-```json
-{ "action": "listPins", "channelId": "C123" }
-```
-
-```json
-{ "action": "memberInfo", "userId": "U123" }
-```
-
-```json
-{ "action": "emojiList" }
-```
-
-## Safety
-
-- Confirm destructive deletes when context is unclear.
-- Keep outbound messages short; avoid Markdown tables.
-- Prefer thread/message IDs over fuzzy channel names.
+Slack formatting, mentions, tables, charts, and interactive controls are described automatically by the current `message` tool hints. Follow those hints instead of maintaining a second action or formatting catalog here.
diff --git a/extensions/tlon/package.json b/extensions/tlon/package.json
index 60ee8ab18da1..b91aebfc0bbf 100644
--- a/extensions/tlon/package.json
+++ b/extensions/tlon/package.json
@@ -10,7 +10,7 @@
   "dependencies": {
     "@aws-sdk/client-s3": "3.1095.0",
     "@aws-sdk/s3-request-presigner": "3.1095.0",
-    "@tloncorp/tlon-skill": "0.4.4",
+    "@tloncorp/tlon-skill": "0.4.5",
     "@urbit/aura": "3.0.0",
     "zod": "4.4.3"
   },
diff --git a/extensions/whatsapp/skills/wacli/SKILL.md b/extensions/whatsapp/skills/wacli/SKILL.md
index fffc28e99480..b749dcc1f0a8 100644
--- a/extensions/whatsapp/skills/wacli/SKILL.md
+++ b/extensions/whatsapp/skills/wacli/SKILL.md
@@ -20,7 +20,7 @@ metadata:
             {
               "id": "go",
               "kind": "go",
-              "module": "github.com/steipete/wacli/cmd/wacli@latest",
+              "module": "github.com/openclaw/wacli/cmd/wacli@latest",
               "bins": ["wacli"],
               "label": "Install wacli (go)",
             },
diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index 274fbb6e9425..234b41d00a6d 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -1907,8 +1907,8 @@ importers:
         specifier: 3.1095.0
         version: 3.1095.0
       '@tloncorp/tlon-skill':
-        specifier: 0.4.4
-        version: 0.4.4
+        specifier: 0.4.5
+        version: 0.4.5
       '@urbit/aura':
         specifier: 3.0.0
         version: 3.0.0
@@ -5226,32 +5226,32 @@ packages:
     resolution: {integrity: sha512-5Kc5CM2Ysn3vTTArBs2vESUt0AQiWZA86yc1TI3B+lxXmtEq133C1nxXNOgnzhrivdPZIh3zLj5gDnZjoLL5GA==}
     engines: {node: '>=12.17.0'}
 
-  '@tloncorp/tlon-skill-darwin-arm64@0.4.4':
-    resolution: {integrity: sha512-yr7Lg4KJPWInBFarXhJV+Z8LxFFLGJAgoBY6d6sc7zVW40oKUM+OQjKomKaNy8H4aYk0CnZkXlN+CTOn3/LD6g==}
+  '@tloncorp/tlon-skill-darwin-arm64@0.4.5':
+    resolution: {integrity: sha512-urM3HnsoQR/yqT8qxs4DZu/B6bWS0EdNKcJXme5FdTsuX4NfOhbMAdjl+aasu2G86GF3q/fgH00g75DNiIFB9w==}
     cpu: [arm64]
     os: [darwin]
     hasBin: true
 
-  '@tloncorp/tlon-skill-darwin-x64@0.4.4':
-    resolution: {integrity: sha512-h9yL4fwkLa2zXqROWV6QuqVB4FlQOA9iQfahST15MAjyQRBqJEt2zxjuXHbkvlZ9SKm4vS5J7YLXI6yXLvULGA==}
+  '@tloncorp/tlon-skill-darwin-x64@0.4.5':
+    resolution: {integrity: sha512-w+pr1NwiD6524WAIHwpQ0IjtqV1K7SUw2aqaG27lvluPukzolYWl3fcliJzY2AOeCwe2r3JkCq5JD1oiZGkLjQ==}
     cpu: [x64]
     os: [darwin]
     hasBin: true
 
-  '@tloncorp/tlon-skill-linux-arm64@0.4.4':
-    resolution: {integrity: sha512-OBzY/ACHo9/wecN0XTkepK2mGwDnqcs+T7vEhjHPKr4aD+YGBvdgA0abeXcSsz9EXKVnRnaDMo9QDYcROy4DIw==}
+  '@tloncorp/tlon-skill-linux-arm64@0.4.5':
+    resolution: {integrity: sha512-t/N4nUF2abzQA22ONbjSdOEyQ0k6nWaA/wZ3OZ0Gzmsfa4caEbMwiHS/nY9po9p3Gb8IFXW7ZTMwMe7w2E3XlA==}
     cpu: [arm64]
     os: [linux]
     hasBin: true
 
-  '@tloncorp/tlon-skill-linux-x64@0.4.4':
-    resolution: {integrity: sha512-KCSfVFDBvR/oaDoa/C4Zyvw8eZTq0GDEcaWfn0tbLF3oWDNziSG1a/VFsQzDhHQWI4n0oq2402/HEbMMAI0Fvw==}
+  '@tloncorp/tlon-skill-linux-x64@0.4.5':
+    resolution: {integrity: sha512-1u5cdmnjSeYGhgTSMTDrpenPOxCu7B5Q62cuviBeYaS7lvvNTNwynIwuO9QR3Q98r7rMMyzWExhDHJ+sRlSbsw==}
     cpu: [x64]
     os: [linux]
     hasBin: true
 
-  '@tloncorp/tlon-skill@0.4.4':
-    resolution: {integrity: sha512-kjFEqZNieQZR7aSgdM54GRWfMYtukuZlMzlW6ThNWKwrX2YDl8ieOZ7dA0gGcp2TpPHbtTpj83Ro1DMPlcGZ5A==}
+  '@tloncorp/tlon-skill@0.4.5':
+    resolution: {integrity: sha512-rsQ4Qqm1oCGLz7rW/TFyVc+8KANifGNrY8jTEu4AbIgFfGR9m3I8tdRjSdAkVW7ZCZ7VDI/RlZhCN8s/dNz+Kw==}
     hasBin: true
 
   '@tokenizer/inflate@0.4.1':
@@ -12387,24 +12387,24 @@ snapshots:
   '@tinyhttp/content-disposition@2.2.4':
     optional: true
 
-  '@tloncorp/tlon-skill-darwin-arm64@0.4.4':
+  '@tloncorp/tlon-skill-darwin-arm64@0.4.5':
     optional: true
 
-  '@tloncorp/tlon-skill-darwin-x64@0.4.4':
+  '@tloncorp/tlon-skill-darwin-x64@0.4.5':
     optional: true
 
-  '@tloncorp/tlon-skill-linux-arm64@0.4.4':
+  '@tloncorp/tlon-skill-linux-arm64@0.4.5':
     optional: true
 
-  '@tloncorp/tlon-skill-linux-x64@0.4.4':
+  '@tloncorp/tlon-skill-linux-x64@0.4.5':
     optional: true
 
-  '@tloncorp/tlon-skill@0.4.4':
+  '@tloncorp/tlon-skill@0.4.5':
     optionalDependencies:
-      '@tloncorp/tlon-skill-darwin-arm64': 0.4.4
-      '@tloncorp/tlon-skill-darwin-x64': 0.4.4
-      '@tloncorp/tlon-skill-linux-arm64': 0.4.4
-      '@tloncorp/tlon-skill-linux-x64': 0.4.4
+      '@tloncorp/tlon-skill-darwin-arm64': 0.4.5
+      '@tloncorp/tlon-skill-darwin-x64': 0.4.5
+      '@tloncorp/tlon-skill-linux-arm64': 0.4.5
+      '@tloncorp/tlon-skill-linux-x64': 0.4.5
 
   '@tokenizer/inflate@0.4.1(supports-color@10.2.2)':
     dependencies:
diff --git a/src/plugins/contracts/channel-provider-skills.contract.test.ts b/src/plugins/contracts/channel-provider-skills.contract.test.ts
new file mode 100644
index 000000000000..d2b5f031d9c9
--- /dev/null
+++ b/src/plugins/contracts/channel-provider-skills.contract.test.ts
@@ -0,0 +1,88 @@
+import { readFileSync } from "node:fs";
+import { dirname, relative, resolve } from "node:path";
+import { describe, expect, it } from "vitest";
+import { listGitTrackedFiles } from "../../test-utils/repo-files.js";
+
+type PluginManifest = {
+  channels?: unknown;
+  skills?: unknown;
+};
+
+const RETIRED_SKILL_PATTERNS = [
+  {
+    pattern: /use\s+the\s+[`'"]slack[`'"]\s+tool/iu,
+    replacement: "the `message` tool",
+  },
+  {
+    pattern:
+      /["']action["']\s*:\s*["'](?:sendMessage|readMessages|editMessage|deleteMessage|pinMessage|unpinMessage|listPins|memberInfo|emojiList)["']/u,
+    replacement: "a canonical `message` action",
+  },
+  { pattern: /owner_open_id/u, replacement: "grant_to_requester" },
+  { pattern: /feishu_doc_list_blocks/u, replacement: "feishu_doc action=list_blocks" },
+  { pattern: /feishu_doc_update_block/u, replacement: "feishu_doc action=update_block" },
+  { pattern: /feishu_doc_delete_block/u, replacement: "feishu_doc action=delete_block" },
+  {
+    pattern: /["']components["']\s*:\s*["']\[Carbon v2 components\]["']/u,
+    replacement: "structured components",
+  },
+  { pattern: /## 备用方案（直接使用 `cron` 工具）/u, replacement: "qqbot_remind only" },
+  {
+    pattern: /github\.com\/steipete\/wacli\/cmd\/wacli@latest/u,
+    replacement: "github.com/openclaw/wacli/cmd/wacli@latest",
+  },
+] as const;
+
+function listRepositoryOwnedChannelSkillFiles(): string[] {
+  const trackedFiles = listGitTrackedFiles({ pathspecs: ["extensions"] }) ?? [];
+  const trackedFileSet = new Set(trackedFiles);
+  const skillFiles = new Set<string>();
+
+  for (const manifestPath of trackedFiles.filter((file) =>
+    /^extensions\/[^/]+\/openclaw\.plugin\.json$/u.test(file),
+  )) {
+    const manifest = JSON.parse(
+      readFileSync(resolve(process.cwd(), manifestPath), "utf8"),
+    ) as PluginManifest;
+    if (!Array.isArray(manifest.channels) || !Array.isArray(manifest.skills)) {
+      continue;
+    }
+
+    const pluginDir = dirname(manifestPath);
+    for (const skillRoot of manifest.skills) {
+      if (typeof skillRoot !== "string" || skillRoot.includes("node_modules")) {
+        continue;
+      }
+      const relativeRoot = relative(process.cwd(), resolve(pluginDir, skillRoot)).replaceAll(
+        "\\",
+        "/",
+      );
+      for (const file of trackedFileSet) {
+        if (file.startsWith(`${relativeRoot}/`) && file.endsWith(".md")) {
+          skillFiles.add(file);
+        }
+      }
+    }
+  }
+
+  return [...skillFiles].toSorted();
+}
+
+describe("bundled channel-provider skill contracts", () => {
+  it("does not teach retired tool, action, parameter, or install contracts", () => {
+    const failures: string[] = [];
+    const skillFiles = listRepositoryOwnedChannelSkillFiles();
+
+    expect(skillFiles.length).toBeGreaterThan(0);
+    for (const file of skillFiles) {
+      const source = readFileSync(resolve(process.cwd(), file), "utf8");
+      for (const entry of RETIRED_SKILL_PATTERNS) {
+        if (entry.pattern.test(source)) {
+          failures.push(`${file}: replace ${entry.pattern.source} with ${entry.replacement}`);
+        }
+      }
+    }
+
+    expect(failures).toStrictEqual([]);
+  });
+});