Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/nongjun/feishu-cursor-claw/agents-mdgit clone --depth 1 https://github.com/nongjun/feishu-cursor-clawWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/nongjun/feishu-cursor-claw/agents-md)<a href="https://agentmods.dev/instructions/nongjun/feishu-cursor-claw/agents-md"><img src="https://agentmods.dev/badge/instructions/nongjun/feishu-cursor-claw/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01840 | $0.01840 |
| Opus 5 | $0.00920 | $0.00920 |
| Sonnet 5 | $0.00368 | $0.00368 |
| Haiku 4.5 | $0.00184 | $0.00184 |
Grade A, and why
feishu-cursor-claw AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — feishu-cursor-claw(虾群平台版)
本文件是 AI 编码助手理解本项目的入口。所有信息矛盾时以
文档/目录为准。
项目定位
飞书 → Cursor AI 远程遥控桥接服务。用户在飞书发消息,server 自动转发给本地 Cursor Agent CLI 执行,执行结果通过飞书卡片回传。灵感来自 OpenClaw。
技术栈
| 层 | 技术 |
|---|---|
| 运行时 | Bun 1.x + TypeScript(直接运行,无需编译) |
| 飞书 SDK | @larksuiteoapi/node-sdk(WebSocket 长连接) |
| 数据库 | SQLite(Bun 内置,记忆向量索引 + FTS5) |
| 语音 | 火山引擎豆包 STT → 本地 whisper-cpp 兜底 |
| 向量 | 火山引擎豆包 Embedding API |
| 部署 | macOS launchd(service.sh 管理) |
目录结构
虾群平台版/
├── server.ts # 主服务入口:飞书 WebSocket → Cursor Agent CLI
├── bridge.ts # OpenAI API 桥接(供 OpenClaw 调用)
├── memory.ts # 记忆管理器 v2(SQLite + 向量 + FTS5)
├── memory-tool.ts # 记忆 CLI(Cursor Agent 通过 shell 调用)
├── scheduler.ts # 定时任务调度器(cron-jobs.json 驱动)
├── heartbeat.ts # 心跳系统(定期触发 HEARTBEAT.md 检查)
├── sync-apple-notes.ts # Apple Notes 同步
├── backfill-embeddings.ts # 向量嵌入回填工具
├── feishu/ # 飞书集成(17 个 .ts 文件)
│ ├── client.ts # 飞书 SDK 客户端
│ ├── send.ts # 消息发送
│ ├── streaming-card.ts # 流式卡片更新
│ ├── mention.ts # @提及处理
│ └── types.ts # 类型定义
├── templates/ # 工作区初始化模板
├── plugins/ # 插件系统
│ └── turix-cua/ # 桌面操控代理(AI 视觉 + 鼠标键盘操控)
├── 瑞小美-harness/ # AI 编程辅助配置工具包
├── .cursor/ # 本项目 Cursor 配置
├── 参考代码/ # 参考项目集合(不参与构建)
├── 文档/ # 结构化知识库(唯一真相源)
├── 架构.md # 系统架构鸟瞰图
├── .env / .env.example # 环境变量
├── package.json # 项目依赖
└── service.sh / setup.sh # 部署脚本
模块职责与依赖
进程模型
单进程架构,server.ts 是唯一的进程入口。所有模块在同一进程内运行。
模块关系
- server.ts — 主进程,实例化 memory.ts、scheduler.ts、heartbeat.ts,导入 feishu/ 模块
- memory.ts — 被 server.ts 实例化,提供记忆读写能力
- memory-tool.ts — 独立 CLI 进程,供 Cursor Agent 通过 shell 调用记忆系统
- scheduler.ts — 被 server.ts 实例化,读取 cron-jobs.json 驱动定时任务
- heartbeat.ts — 被 server.ts 实例化,定期触发 HEARTBEAT.md 检查
- bridge.ts — 独立服务,提供 OpenAI 兼容 API(供 OpenClaw 调用)
- feishu/ — 飞书功能集合,被 server.ts 通过 SDK 导入
- plugins/turix-cua/ — 独立 Python 进程,通过
turix.sh启动,AI 截屏→理解→操控鼠标键盘完成桌面任务
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 150 lines · 1,840 tokens per session scan A 48f9639ccdce
feishu-cursor-claw AGENTS.md is an instructions file published in the GitHub repository nongjun/feishu-cursor-claw (14 stars, last pushed 1mo ago), licensed MIT. It adds 1,840 tokens to every session, about $0.0092 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
metabot CLAUDE.md
Claude Code instructions for xvirobotics/metabot: The canonical project instructions are in AGENTS.md. Follow that file for this repository. This compatibility file exists for Claude Code; MetaBot install/update does not deploy or replace workspace instruction files.
cursor-handbook CLAUDE.md
Claude Code instructions for girijashankarj/cursor-handbook, covering claude.md — instructions for claude, project, critical rules, token efficiency (most important) and security (critical).
fast-mcp-telegram CLAUDE.md
Claude Code instructions for leshchenko1979/fast-mcp-telegram, covering fast-mcp-telegram, session corrections and 2026-05-27.
meta-agent-skills AGENTS.md
AGENTS.md instructions for yu-iskw/meta-agent-skills, covering agents.md, available skills, available agents, agent skills and what are agent skills?.
openlark CLAUDE.md
Claude Code instructions for foxzool/openlark, a project described as: 飞书开放平台的非官方 Rust SDK,支持自定义机器人、长连接机器人、云文档、飞书卡片、消息、群组等 API 调用。.
seedream-image-mcp CLAUDE.md
Instructions for wearzdk/seedream-image-mcp, covering apis, testing and frontend.