# Operate Live Lark Meetings

Live meeting automation requires careful identity handling and separate identifiers. This skill guides authorized Lark discovery, participation, event reading, and communication through lark-cli.

## Install

```bash
npx skillstore add larksuite/lark-vc-agent
```

## Metadata

- Status: approved
- Slug: larksuite-lark-vc-agent
- Version: 1.0.0
- Author version: 1.0.0
- Skillstore revision: r2
- Version status: valid
- Tree hash: 242b1643fab7e8dc471001853f64662dc7750b88591ada465da1bdd2f4ade362
- Author: larksuite
- GitHub username: larksuite
- License: MIT
- Repository: https://github.com/larksuite/cli/tree/main/skills/lark-vc-agent/
- Ref: ebdfe608f5de2b66ff37ab4af12af8ac4f5e8006
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: network, filesystem, external\_commands
- Quality score: 78
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/larksuite-lark-vc-agent
- Manifest: https://skillstore.pages.dev/api/skills/larksuite-lark-vc-agent/manifest

## Capabilities

- Discover active meetings visible to the current user or an authorized application bot.
- Join an active meeting as an application bot using a nine-digit meeting number.
- Read paginated participant, chat, transcript, and content-sharing events from a live meeting.
- Send authorized in-meeting text messages, reaction emojis, and video meeting feedback.
- Leave a meeting as the application bot after an explicit user request.

## Use Cases

- Deploy a Meeting Assistant: Join a confirmed live meeting, record its long meeting ID, and monitor events without leaving until requested.
- Monitor and Respond Live: Review current speakers, participant changes, chat, transcripts, and shared content before sending an approved message or reaction.
- Discover Authorized Meetings: Find active meetings visible to a user or application identity and preserve that identity for later event access.

## Prompt Templates

### Find My Active Meetings

```
Find the active Lark meetings visible to my current user identity. Show each title, meeting number, and long meeting ID.
```

### Send Meeting Feedback

```
In my current meeting, send the reaction that means [feedback]. Confirm the selected meeting and reaction before sending.
```

### Join and Monitor

```
Join meeting [nine-digit number] as the application bot after confirming the number. Then retrieve all currently visible events and summarize them.
```

### Analyze an Authorized Meeting

```
Using application identity, discover meetings for [user open ID], let me choose one, fetch every event page, and summarize recent discussion and shared content.
```

## Limitations

- The feature is in limited beta and requires eligible Lark accounts, application permissions, installation, and configured data scope.
- It does not search ended meetings or retrieve participant snapshots, minutes, transcripts, or recordings.
- Application event reading and messaging require the bot to be present or previously present within the supported grace window.
- Operations require lark-cli, valid authentication, and correct handling of meeting numbers and long meeting IDs.

## Best Practices

- Confirm a nine-digit meeting number and its source before visible bot joins. Use dry run when identity or parameters are uncertain.
- Keep the identity path consistent, and distinguish the nine-digit meeting number from the long meeting ID.
- Fetch current events with full pagination before live-state answers, and request explicit approval for every visible write action.

## Anti Patterns

- Do not treat any nine-digit number as automatic permission to join a meeting.
- Do not reuse stale or partial event results when answering current meeting questions.
- Do not use this skill for ended meetings, participant snapshots, recordings, minutes, transcripts, or bound group chat messages.

## Security Audit

- Audited at: 2026-07-23T17:25:06.622\+00:00
- Summary: All 208 static findings are false positives caused by Markdown links, inline code, examples, Chinese text, and entropy heuristics. No path traversal, reconnaissance, obfuscation, code execution, or automatic network request is present. Documented Lark actions require authenticated identities and explicit user intent for visible write operations.

## Stats

- Views: 32
- Downloads: 7
- Favorites: 0
- Popularity score: 0