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/deepcoldy/botmux/claude-mdgit clone --depth 1 https://github.com/deepcoldy/botmuxWhat 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 | $0.04000 | $0.04000 |
| Opus 5 | $0.02000 | $0.02000 |
| Sonnet 5 | $0.00800 | $0.00800 |
| Haiku 4.5 | $0.00400 | $0.00400 |
Grade A, and why
botmux CLAUDE.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 3d 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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
botmux
飞书话题群 ↔ AI 编程 CLI 桥接。Daemon 监听飞书消息，每个新话题自动 spawn 一个独立 CLI 进程（Claude Code / Codex / Gemini 等 20+ 种，完整列表见 README）。
构建 & 运行
bun run build # tsc 编译
bun run daemon:restart # 重启 daemon（自动恢复 active sessions）
bun run daemon:logs # 查看日志
- 每次修改后需要
bun run build然后bun run daemon:restart
包管理器是 bun（packageManager: [email protected]，锁文件 bun.lock）。装依赖用 bun install --frozen-lockfile。
⚠️ trustedDependencies: ["electron","node-pty"] 不能删：bun 默认不跑依赖的生命周期脚本，而 node-pty 要靠它 node-gyp 编出 build/Release/pty.node —— 少了这个，PTY 全废、编译版二进制也打不出来（pty.node 是被嵌进去的）。electron 的 postinstall 负责下载对应平台的二进制。
这个名单刻意只有两项（与 pnpm 时代的 onlyBuiltDependencies 逐字一致），别照着"顺手补全"往里加 esbuild —— 实测 esbuild 虽然有 postinstall，但它的二进制由 @esbuild/<platform> 平台包直接提供：空白目录里 bun install esbuild 不跑任何脚本，esbuild --version 照样输出 0.28.2。加进白名单只会让 bun 比 pnpm 多跑脚本、扩大两者的行为差异，与迁移目标相反。
注意与「用户怎么装 botmux」区分开：install-diagnostics.ts 的 InstallKind（含 'pnpm-global'）与 maintenance.ts 的自动更新说的是终端用户的安装方式，线上确实有人 pnpm i -g botmux。那些不是本仓库的构建工具链，不要跟着一起改。
Bun 开发链路
daemon / supervisor / dashboard 都能直接跑 TypeScript，不必先 bun run build：
bun run daemon:bun # bun src/index-daemon.ts
bun run supervisor:bun # bun src/index-supervisor.ts
bun run dashboard:bun # bun src/index-dashboard.ts
bun run build:bun # 打自包含单文件二进制（scripts/build-bun-binary.mjs）
发版编译用的 Bun 版本钉在 1.4.0（见 .github/workflows/）。本地 bun 与它差太多时，编译产物的行为可能和 CI 不一致——排查编译态问题前先核对 bun --version。
测试里 spawn 子进程必须走 test/helpers/ts-runner.ts，不要写 spawn(process.execPath, ['--import','tsx', …])：那是 Node-only 形态，Bun 下 process.execPath 是 bun 二进制、bun --import tsx 不合法，子进程会全部起不来。helper 按运行时解析（Node 加 tsx loader、Bun 原生跑 TS）。片段里 import 仓库模块（.js specifier 实际是 .ts）时用 spawnTsEvalWithRepoImports，普通 spawnTsEval 在 Node 下会 ERR_MODULE_NOT_FOUND。
worktree 里也能直接用 bun，且不额外占依赖体积——只要不跑 install（见下节）。实测在「源码 + node_modules symlink 到 canonical」的 worktree 里，bun src/cli.ts、bun run build、tsc、vitest、build:bun（pty.node 经 symlink 解析，不复制）以及编出二进制跑 smoke 六项，全部正常。
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.
- 3d ago First seen · 138 lines · 4,000 tokens per session scan A e96e2572ac38
botmux CLAUDE.md is an instructions file published in the GitHub repository deepcoldy/botmux (1,264 stars, last pushed 3d ago), licensed MIT. It adds 4,000 tokens to every session, about $0.0200 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
openlark AGENTS.md
Instructions for foxzool/openlark, covering openlark knowledge base, structure, where to look, conventions and 模块组织.
larksnap CLAUDE.md
Instructions for AmbroseX/larksnap, covering 基本约定, plan 文件管理, 开发规范, 根据场景动态调整顺序 and 文档写入.
Claude-to-IM-skill CLAUDE.md
Instructions for op7418/Claude-to-IM-skill, covering claude.md — project guidelines for claude-to-im-skill and replying to github issues.
lark-mcp-on-agentcore AGENTS.md
AGENTS.md instructions for ddpie/lark-mcp-on-agentcore, covering agents.md, project overview, setup & commands, project structure and code style.
lodestar AGENTS.md
Instructions for leviyuan/lodestar, covering lodestar 项目指引, 范围地图, 工作规则, 跨模块不变式 and live daemon 操作.
cc-connect-next AGENTS.md
AGENTS.md instructions for timmyagentic/cc-connect-next, covering cc-connect development guide, project overview, architecture, key design principles and core interfaces.