# Manage Lark Messages and Chats

Teams need controlled access to Lark messages, chats, files, and cards from AI tools. This skill guides Claude, Codex, and Claude Code through lark-cli workflows with clear permission boundaries.

## Install

```bash
npx skillstore add larksuite/lark-im
```

## Metadata

- Status: approved
- Slug: larksuite-lark-im
- Version: 1.0.0
- Author version: 1.0.0
- Skillstore revision: r1
- Version status: valid
- Tree hash: ff45d40266eb326b2786ae88185b02def439a3e5a396e49897593802e6757c9e
- Author: larksuite
- GitHub username: larksuite
- License: MIT
- Repository: https://github.com/larksuite/cli/tree/main/skills/lark-im/
- 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: 80
- Quality tier: silver
- Public page: https://skillstore.pages.dev/skills/larksuite-lark-im
- Manifest: https://skillstore.pages.dev/api/skills/larksuite-lark-im/manifest

## Capabilities

- Send and reply to Lark messages using text, Markdown, post content, media, or interactive cards.
- Search and list chats, messages, threads, members, reactions, and message metadata through documented lark-cli commands.
- Download image and file resources from messages with relative output paths and chunked download support.
- Create or update chats, resolve chat identifiers, and handle user or bot identity differences.
- Manage message flags, feed shortcuts, feed groups, reactions, and card action callback workflows.

## Use Cases

- Coordinate Team Updates: Send approved status updates, replies, files, and images to selected Lark chats or direct messages.
- Review Chat History: Search messages by keyword, sender, chat, attachment type, and time range before producing summaries.
- Build Interactive Workflows: Create Lark interactive cards and connect card button callbacks to documented event handling steps.

## Prompt Templates

### Send a Simple Message

```
Send this approved message to the Lark chat named <chat name>: <message>. Confirm the recipient and sending identity before running commands.
```

### Find Recent Messages

```
Search Lark messages for <keyword> in <chat or scope> from <start date> to <end date>. Return sender, time, chat, and message identifiers.
```

### Download Message Resources

```
Find the message resource for <message description> and download the image or file to a relative output path. Explain any permission errors.
```

### Create an Interactive Card Flow

```
Design an interactive Lark card for <workflow>. Follow the card creation reference, send it only after approval, and outline callback handling.
```

## Limitations

- The user must install and authenticate lark-cli before the workflows can run.
- The skill cannot bypass Lark permissions, scopes, tenant rules, or chat membership checks.
- Some operations only support user identity, while others require bot identity.
- Remote media upload or download behavior depends on network access and Lark API availability.

## Best Practices

- Confirm recipient, content, and sending identity before sending any message.
- Use narrow filters and pagination when searching message history for summaries.
- Prefer dry runs and relative paths when testing sends, uploads, downloads, and card payloads.

## Anti Patterns

- Do not send messages or urgent notifications without explicit user approval.
- Do not assume a bot can access chats, users, files, or sender names without the required scopes.
- Do not paste unreviewed card payloads or local file paths into message content.

## Security Audit

- Audited at: 2026-07-23T16:24:06.203\+00:00
- Summary: The package contains Markdown documentation and command references for Lark IM. The static findings are false positives from Markdown links, placeholder URLs, inline code formatting, and documented lark-cli examples. No evidence found of hidden executable files, encoded payloads, prompt injection, or malicious data exfiltration intent.

## Stats

- Views: 96
- Downloads: 13
- Favorites: 0
- Popularity score: 0