Agent skill

Managed Lark Docs

by huangruiteng in huangruiteng/CS-Notes

Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits.

MITAuto-check passedProductivity & Automation

Install Managed Lark Docs

skills CLI
$ npx skills add huangruiteng/CS-Notes --skill managed-lark-docs -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install huangruiteng/CS-Notes managed-lark-docs --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/huangruiteng/CS-Notes.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.codex/skills/managed-lark-docs .claude/skills/managed-lark-docs && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
managed-lark-docs
GitHub stars
4k
Token cost
~3.1k tokens
SKILL.md length
1,624 words
Files
4 (incl. references)
Skills in repo
39
Repo updated
First seen
Licence
MIT

At a glance

Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits.

  • Works in 6 steps: Establish Freshness → Plan The Change → Update The Local Review Surface → …
  • Iterative document updates
  • SKILL.md covers Core Model, Non-Negotiable Defaults, Required Skills and Project And Authority Resolution, plus 5 more sections
  • Calls git

What it does

Managed Lark Docs is an agent skill from huangruiteng/CS-Notes. Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits. Use for iterative document updates, local-mirror sync, or durable project ownership; LoopX integration is optional.

Its SKILL.md is about 3.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files, including reference files (for example `agents/openai.yaml`, `references/comment-preserving-sync.md` and `references/loopx-integration.md`).

It sits in Productivity & Automation, covering Messaging and chat bots. It works with Feishu (Lark) and Git. The licence is MIT.

When your agent uses it

  • Iterative document updates
  • Local-mirror sync
  • Durable project ownership
  • LoopX integration is optional

Example prompts

  • “/managed-lark-docs”

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. Establish Freshness
  2. Plan The Change
  3. Update The Local Review Surface
  4. Respect The Remote Write Gate
  5. Apply The Narrowest Safe Write
  6. Verify And Write Back

What it can do on your machine

Read from SKILL.md and the folder at commit f7b4e92. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Managed Lark Docs loads about 3.1k tokens when it runs, and up to ~4.2k if it reads all its reference files. Until then it costs about 64 tokens; SKILL.md has 1,624 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~64
When it runs · the whole SKILL.md, loaded when a task matches
~3.1k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~4.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from huangruiteng/CS-Notes at commit f7b4e92, republished under its MIT licence (© huangruiteng). 1,624 words, ~3,116 tokens.

Download SKILL.mdSave it as .claude/skills/managed-lark-docs/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
managed-lark-docs
description
Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits. Use for iterative document updates, local-mirror sync, or durable project ownership; LoopX integration is optional.

Managed Lark Docs

This skill manages a Lark document as a durable project material, not as a one-off text-editing target. It is repository-neutral and works without LoopX. Project-specific document conventions come from the owning project.

Core Model

A managed Lark document has four distinct surfaces:

  1. Remote source: the Lark doc and its current revision, comments, rich-text marks, cites, title, and outline.
  2. Project authority: the owning project's document policy and, when already connected to LoopX, its goal and redacted authority source registration.
  3. Local working state: an ignored mirror and manifest under the owning project's .local/ directory.
  4. Publication surface: any intentionally tracked, shared, or public derivative. This is opt-in and must pass the project's privacy boundary.

Do not collapse these surfaces. In particular, a local mirror is not automatically a Git artifact, and a registry entry stores lineage and authority, not the private body or raw source URL.

Non-Negotiable Defaults

  • Resolve the owning project before creating a mirror. Resolve a LoopX goal_id only when the project already uses LoopX for this material.
  • When the project is connected to LoopX, register the source with loopx register-authority-source; do not rely on chat memory or a repository- specific document catalog alone.
  • Store private or internal mirrors under <project-root>/.local/managed-lark-docs/<source-id>/ by default.
  • In a Git project, verify the target path with git check-ignore -v before writing. Stop if it is not ignored. Outside Git, use an owner-approved private directory rather than a published or shared folder.
  • Never put a new private managed document in a tracked docs/ path merely to obtain a diff or registry entry.
  • A Git-tracked mirror is allowed only when the user explicitly asks for one, the repository declares that document family as tracked authority, and the content is safe for that repository's audience.
  • Treat the remote Lark doc as current truth unless the user explicitly names a canonical local source. A local canonical source still does not authorize erasing newer remote edits or reviewer signal.
  • Preserve unresolved comments, highlights, manual emphasis, native cites, and clearer remote-only section boundaries.
  • Use the narrowest safe remote write and fetch again after writing.
  • Never store credentials, access tokens, raw private source bodies, or private operational links in Git or public-safe LoopX projections.

Required Skills

  • Read and follow lark-doc before fetching or updating a document.
  • Use lark-drive for comments, title metadata, permissions, and comment anchor checks.
  • Use lark-shared for identity, authentication, and missing scopes.
  • When the project is connected to LoopX, read and follow loopx-doc-registry and LoopX integration.
  • When editing this skill, use skill-creator and run its validator.

If an official lark-* skill is absent, try the installed CLI lark-cli skills read <skill-or-reference-path> for its bundled instructions. If neither source is available, report the missing prerequisite and hold the affected operation. Do not install tools, change authorization, or invent API parameters merely to make this skill available.

Project And Authority Resolution

Find the owning repository from the user's named project, target document context, or current working directory. Reuse its current document policy and choose a stable, public-safe source_id; keep private names and raw URLs only in the local manifest. Do not pick a project from historical usage.

For a project already connected to LoopX, follow LoopX integration. Otherwise keep the private mirror and manifest without requiring a Goal or registry. Do not bootstrap or connect a project as a side effect of document editing.

Local Layout

Use this default layout:

text
<project-root>/.local/managed-lark-docs/<source-id>/
├── manifest.json
├── mirror.lark.md
└── drafts/                 # only when a separate review draft is useful

The mirror filename may use XML or another loss-aware format when Markdown cannot represent the document faithfully. The manifest should contain only local/private state and use a small contract such as the following. Omit goal_id for projects without LoopX:

json
{
  "schema_version": "managed_lark_doc_local_manifest_v0",
  "source_id": "stable-source-id",
  "goal_id": "owning-goal-id",
  "boundary": "local_private",
  "source_url": "raw private URL allowed only in .local",
  "remote_revision": 12,
  "remote_revision_observed_at": "ISO-8601 timestamp",
  "mirror_path": ".local/managed-lark-docs/stable-source-id/mirror.lark.md",
  "mirror_sha256": "sha256",
  "mirror_status": "remote_readback_verified",
  "git_policy": "ignored_local_only",
  "conflict_rule": "Remote/current-code authority rule"
}

After every remote update, refresh the revision, timestamp, digest, status, and any local supporting-artifact references. Do not copy raw private values from this manifest into LoopX's public-safe status or Git artifacts.

Managed Update Workflow

1. Establish Freshness

Before drafting or writing:

  1. Fetch the current title, revision, and outline.
  2. Fetch only the relevant sections when the request is narrow; fetch the full document when a broad rewrite or structural comparison requires it.
  3. Check unresolved comments and nearby rich-text marks when the change touches reviewed content.
  4. Compare the remote revision and outline with the local manifest and mirror.
  5. Classify remote-only changes as preserve, merge, ask_user, or explicit_replace. Default to preserve or ask_user.

When the remote revision is newer, it becomes the base for that sync. Do not overwrite it from an older local mirror.

2. Plan The Change

Write for the reader, not as a work log:

  • integrate evidence into stable sections;
  • preserve useful original claims and citations;
  • distinguish verified fact, inference, proposal, and open gate;
  • avoid timestamped assistant logs unless the document is itself an execution log;
  • prefer replacing or compressing stale material over endlessly appending;
  • keep private evidence in .local and expose only generalized conclusions on shareable surfaces.

Use the target document's voice by default. Apply a personal style corpus such as CS-Notes only when the user requests it or the document family declares it. Sample enough representative writing to infer structure and rhythm; do not mechanically copy private phrasing.

3. Update The Local Review Surface

Update mirror.lark.md first when the user wants a reviewable local diff. A separate draft is optional and should be created only when:

  • the user asks to see a draft first;
  • the rewrite is broad and remote sync is not yet approved;
  • a standalone memo is the requested artifact; or
  • the mirror format is too loss-aware or noisy for comfortable review.

For private documents, the review surface remains in .local; reviewability does not require Git.

4. Respect The Remote Write Gate

If the user asked only to draft, refine locally, or show a preview, stop before the Lark write. If the current request explicitly says to update/sync the Lark document, that is sufficient intent for the scoped write after freshness and boundary checks.

Summarize the planned sections, protected anchors, and unresolved ambiguity before a broad rewrite. Ask only when the ambiguity affects authorship, authority, privacy, or destructive replacement.

Show full SKILL.md (621 more words)Show less
5. Apply The Narrowest Safe Write

Read the official lark-doc update and format references for the installed CLI. Fetch the affected XML with full detail before a preservation-sensitive edit. Build a small patch plan from the current remote state and revision. Use the observed base revision when the installed API supports it. If a write reports a revision conflict, fetch and merge the new remote state before retrying.

Prefer, in order:

  1. exact inline replacement in an unprotected block;
  2. block insertion around protected content;
  3. block replacement of an unprotected section;
  4. whole-document overwrite only when the document has no protected reviewer signal and the user explicitly approved replacement.

Never use a Markdown overwrite merely because it is convenient. A request to "sync", "polish", or "reorganize" does not authorize clearing the document. Native doc cites should remain cites; comments and marked text are not disposable formatting.

str_replace searches the whole document: verify every match and choose block operations when the text is ambiguous or spans multiple blocks. Read back after each patch round before planning another structural edit; replacement can invalidate block IDs. A successful response is not proof that anchors survived. Inspect partial_success, warnings, and actual remote state before retrying; never replay a partially applied batch blindly.

Read references/comment-preserving-sync.md before changing content with comments, highlights, manual emphasis, tables with review anchors, or native Lark cites.

6. Verify And Write Back

After writing:

  1. Fetch the new revision and affected sections.
  2. Fetch the outline and confirm protected section boundaries still exist.
  3. Recheck unresolved comments and touched anchors where applicable.
  4. Confirm native cites and marked spans survived.
  5. Refresh the ignored mirror and local manifest.
  6. When already integrated with LoopX, update the registered source revision and relevant state through its existing owners. See LoopX integration.

Do not claim synced when only one section was updated while other mirror sections remain local-only. Report a precise partial-sync status instead.

Privacy And Git Decision

SituationDefault locationGit?
Private/internal Lark bodyproject .local/managed-lark-docs/No
Raw URL/token/revision manifestsame ignored local directoryNo
Redacted source identity and authorityproject LoopX registryRegistry only
Generalized reusable conclusionproject docs, after boundary reviewExplicit decision
Existing tracked document familyrepository-declared pathOnly if its profile and audience permit

Before any tracked write, scan for internal actors, private links, raw tokens, credentials, local absolute paths, screenshots, and source bodies. If a private mirror was mistakenly committed, move the live mirror to the owning project's ignored local state, remove the current Git projection through a normal review change, and report that prior Git history still retains it. Never rewrite history without explicit owner approval and coordination.

Writing Guardrails

  • Write as if the document is authored by the user or project, not as if an assistant is explaining its edits.
  • Remove drafting phrases such as 建议改写, 给你 review, 这里应该, and process narration from the final body.
  • Prefer direct claims over correction frames like 不是 X,而是 Y when a positive statement is clearer.
  • Explain abstract systems with one concrete end-to-end or failure case.
  • Keep links and citations near the claims they support.
  • Preserve enough context for a new reader; do not assume prior chat history.
  • Keep evidence boundaries explicit and avoid turning hypotheses into shipped facts.
  • Do a whole-document phrase and privacy scan before remote sync.

Output Standard

Report:

  • owning project and stable source_id;
  • LoopX goal and redacted authority topic only when already integrated;
  • local mirror/manifest paths and proof they are ignored;
  • remote revision before and after;
  • sections changed and protected signal preserved;
  • any partial-sync, access, privacy, or history-retention caveat;
  • tracked repository changes, if any, as a separate explicit surface.

Do not expose raw private URLs, tokens, source bodies, credentials, or local private evidence in public summaries.

© huangruiteng, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in .codex/skills/managed-lark-docs of huangruiteng/CS-Notes.

  • SKILL.md
  • agents/openai.yaml
  • references/comment-preserving-sync.md
  • references/loopx-integration.md

Open the folder on GitHubat commit f7b4e92

Compare with similar skills

Managed Lark Docs next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Managed Lark Docs compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Managed Lark Docs this skillhuangruiteng/CS-Notes4k—~3.1kAutomated safety check: PassMIT
Feishu Doc Scraperdaymade/claude-code-skills1.4k—~6.9kAutomated safety check: PassMIT
Feishu Docopenclaw/openclaw392k—~516Automated safety check: PassMIT
Claude To Imop7418/Claude-to-IM-skill2.9k—~3.4kAutomated safety check: NotesMIT
Feishu Docraucvr/Group-Goki1123 repos~592Automated safety check: PassMIT
Feishu BridgeAlexAnys/feishu-openclaw3171 repos~615Automated safety check: PassNone

Similar skills

  • Feishu Doc Scraper

    daymade/claude-code-skills

    Extracts Feishu/Lark Docs, Wiki, Sheets, and Minutes (妙记) transcripts into faithful local Markdown via the lark-cli API — no LLM paraphrasing, browser-DOM fallback when lark-cli can't reach content.

    1.4k GitHub stars~6.9k tokensUpdated today
    Productivity & AutomationAuto-check passed
  • Feishu Doc

    openclaw/openclaw

    Feishu document read/write workflows. An agent skill from openclaw/openclaw.

    392k GitHub stars~516 tokensUpdated today
    Productivity & AutomationAuto-check passed
  • Claude To Im

    op7418/Claude-to-IM-skill

    Bridge THIS Claude Code or Codex session to Telegram, Discord, Feishu/Lark, QQ, or WeChat so the user can chat with Claude from their phone.

    2.9k GitHub stars~3.4k tokensUpdated 6 mo ago
    Productivity & AutomationAuto-check: notes
  • Feishu Doc

    raucvr/Group-Goki

    Feishu document read/write operations. An agent skill from raucvr/Group-Goki.

    112 GitHub starsUsed in 3 repos~592 tokens
    Productivity & AutomationAuto-check passed
  • Feishu Bridge

    AlexAnys/feishu-openclaw

    Connect a Feishu (Lark) bot to Clawdbot via WebSocket long-connection.

    317 GitHub starsUsed in 1 repo~615 tokens
    Productivity & AutomationAuto-check passed
  • Content Collector

    vigorX777/content-collector-skill

    Collect social media content (X/Twitter, WeChat, Jike, Reddit, etc.) into Feishu bitable.

    240 GitHub stars~2k tokensUpdated 6 mo ago
    Productivity & AutomationAuto-check passed

More from huangruiteng/CS-Notes

All 39 skills in this repo
  • CLI Creator

    huangruiteng/CS-Notes

    Build a composable CLI for Codex from API docs, an OpenAPI spec, existing curl examples, an SDK, a web app, an admin tool, or a local script.

    4k GitHub starsUsed in 2 repos~2.7k tokens
    Auto-check passed
  • Codex Thread Heartbeat

    huangruiteng/CS-Notes

    Inspect and manage guarded Codex App-native or launchd heartbeats for Codex main control threads.

    4k GitHub stars~1.5k tokensUpdated yesterday
    Auto-check passed
  • Slack

    huangruiteng/CS-Notes

    A skill your agent uses when you need to control Slack from Clawdbot via the slack tool, including reacting to messages or pinning/unpinning items in Slack channels or DMs.

    4k GitHub starsUsed in 10 repos~578 tokens
    Auto-check passed
  • Codex Thread Reader

    huangruiteng/CS-Notes

    Locate and read a Codex thread by a codex thread link, thread id, or rollout path across all local CODEXHOME directories (~/.codex, ~/.codex-gpt, ...).

    4k GitHub stars~1.4k tokensUpdated yesterday
    Auto-check passed
  • Research Material Scout

    huangruiteng/CS-Notes

    A skill your agent uses when the user asks Codex to research, find learning materials, process "素材:" links, "请你读" / "精读" a material, build a material radar, or use SenSight-like broad information…

    4k GitHub stars~8.3k tokensUpdated yesterday
    Auto-check passed
  • GitHub

    huangruiteng/CS-Notes

    Interact with GitHub using the gh CLI. An agent skill from huangruiteng/CS-Notes.

    4k GitHub starsUsed in 27 repos~279 tokens
    Auto-check passed

Questions about Managed Lark Docs

What does Managed Lark Docs do?

Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits. Managed Lark Docs is an agent skill from huangruiteng/CS-Notes. Maintain and patch existing Lark/Feishu documents while preserving comments, highlights, citations, and newer remote edits.

When should I use Managed Lark Docs?

Managed Lark Docs fits situations like: iterative document updates; local-mirror sync; durable project ownership; loopX integration is optional.

How do I install Managed Lark Docs in Claude Code?

Run `npx skills add huangruiteng/CS-Notes --skill managed-lark-docs -a claude-code`. Or copy the skill folder (.codex/skills/managed-lark-docs in huangruiteng/CS-Notes) into .claude/skills/managed-lark-docs in your project. Claude Code loads it when a task matches its description.

How do I install Managed Lark Docs in Codex?

Run `npx skills add huangruiteng/CS-Notes --skill managed-lark-docs -a codex`. Or copy the skill folder (.codex/skills/managed-lark-docs in huangruiteng/CS-Notes) into .agents/skills/managed-lark-docs in your project. Codex loads it when a task matches its description.

Can I use Managed Lark Docs in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add huangruiteng/CS-Notes --skill managed-lark-docs -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/managed-lark-docs, .gemini/skills/managed-lark-docs, .github/skills/managed-lark-docs and .opencode/skills/managed-lark-docs in your project.

What does Managed Lark Docs need to run?

Going by SKILL.md and its folder, Managed Lark Docs needs the command-line tools its instructions call (git).

Does Managed Lark Docs access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Managed Lark Docs safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Managed Lark Docs use?

Managed Lark Docs is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Managed Lark Docs use?

About 3.1k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 1k tokens, read only when the agent opens those files.

What are the alternatives to Managed Lark Docs?

Skills that share tags, products or a category with Managed Lark Docs: Feishu Doc Scraper (daymade/claude-code-skills, 1.4k stars), Feishu Doc (openclaw/openclaw, 392k stars), Claude To Im (op7418/Claude-to-IM-skill, 2.9k stars) and Feishu Doc (raucvr/Group-Goki, 112 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Managed Lark Docs?

huangruiteng (a GitHub user) maintains it in huangruiteng/CS-Notes, which has 4,000 GitHub stars. The repository holds 39 skills in this directory. The repository was last updated on October 8, 2026.

Source: huangruiteng/CS-Notes on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.