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/alice-jin-dev/bodybridge/claude-mdgit clone --depth 1 https://github.com/alice-jin-dev/bodybridgeWrote 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/alice-jin-dev/bodybridge/claude-md)<a href="https://agentmods.dev/instructions/alice-jin-dev/bodybridge/claude-md"><img src="https://agentmods.dev/badge/instructions/alice-jin-dev/bodybridge/claude-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 | $0.00932 | $0.00932 |
| Opus 5 | $0.00466 | $0.00466 |
| Sonnet 5 | $0.00186 | $0.00186 |
| Haiku 4.5 | $0.00093 | $0.00093 |
Grade A, and why
bodybridge 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 4d 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.
What it actually says
bodybridge — 项目须知（给 Claude Code）
这是什么
让具身设备（StackChan / 树莓派 / ESP32 等）经 claude.ai 自定义连接器接入 AI 的开源云端桥。 V1 的落地样本是 ESP32 基础学习套件（firmware/ 下只有 esp32-bodybridge）；StackChan（小智固件） 是 V2 的样本，不是 V1 要接的设备。核心价值：零 API 费、零电脑依赖、全私有。
四层架构
- MCP Server 层：streamable-http 对外暴露，声明工具清单，接住 Claude 的工具调用。
- Auth 鉴权守门层：完整 OAuth 2.1 —— DCR 无状态注册 + PKCE + JWT 签发 + 中间件逐请求验 JWT。 密钥全走环境变量，绝不写死。 注：最初设计为静态 Token（理由：一人部署一份，不是多用户平台）；后因 claude.ai 自定义 连接器只认 OAuth 入口、不接受静态 token，V1 落地为 OAuth 2.1。详见 MIGRATION.md。
- 设备 Adapter 插槽层（核心）：标准接口 send_command / get_status / list_capabilities。 当前实例是 WebSocketAdapter（adapters/websocket.py）；ESP32 参考固件是第一台用它的设备。 换设备＝实现同一接口，桥身不动。 架构理念：依赖倒置 / 适配器模式，把变化关进插槽。
- Reflex 脊髓反射层（可选）：设备端即时反应，不阻塞 V1。
技术栈
- 语言：Python 3.10+
- 项目管理：uv
- MCP：官方 SDK（mcp），FastMCP，transport = streamable-http
- 无状态优先（stateless_http）
六条血泪铁律（写代码时必须遵守）
- 环境变量向后兼容，绝不乱改名；改名要么保留旧名、要么给清晰迁移映射。
- 每个版本附"迁移必读"清单。
- 输入必须防御性处理，一个字符都不能崩掉服务；最坏结果只能是友好错误提示。
- 错误信息要说人话、暴露原因；不要用 raise_for_status 吞掉真实错误只甩状态码。
- 配置要有合理默认值 + 明确的必填提示。
- 别把 API 版本号等易变参数写死，要可配置。 附：给用户免费/低成本选项；密钥泄露即作废重发。
核心原则
- 桥身求薄：不存业务数据、不做 AI 推理、不碰数据库。越轻越稳。
- 手上只做 V1：只交付 V1 该有的东西，别提前造。
- 一次一个变量：每层跑通了再动下一层，别几层一起改。
协作方式
- 瑾儿 = 总军师（决策、审计、把关每一步）。
- 你（Claude Code）= 施工队，负责执行。
- 先给方案，不要擅自执行：任何改动先说清楚"打算怎么做"，等瑾儿明确指令后再动手。改动先给 diff / 说明，确认后再落地。
- 遇到以下情况必须上报、和瑾儿讨论，不要闷头处理：
- 发现和这份背景冲突，或有更好的做法；
- 指令或需求有模糊不清、可以有多种理解的地方；
- 执行中发现新的可行性、风险或坑。
- 一次一个动作，一次一个变量。绝不硬编码密钥。
说明
- 这份文件里的六条铁律、架构等，是当前阶段的约定，会随项目演进增删。以约定为准，但发现更优解随时提。
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.
- 4d ago First seen · 52 lines · 932 tokens per session scan A f76c14a2a76b
bodybridge CLAUDE.md is an instructions file published in the GitHub repository alice-jin-dev/bodybridge (0 stars, last pushed 24d ago), licensed MIT. It adds 932 tokens to every session, about $0.0047 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 2026-08-31.
Other instructions, from other repositories
cml-mcp AGENTS.md
AGENTS.md instructions for xorrkaz/cml-mcp, covering agents.md — cml mcp server, project overview, compatibility goal, repository layout and tool modules (src/cmlmcp/tools/).
brilliant_sdk AGENTS.md
Instructions for brilliantlabsAR/brilliant_sdk, covering brilliant sdk — agent guide, how an app works (the pattern behind everything), minimal reading paths, verify without hardware and testing.
NeoMind CLAUDE.md
Claude Code instructions for camthink-ai/NeoMind, covering neomind — edge ai platform for iot, development commands, ecosystem repositories, extension package contract (.nep) and device type template contract (json).
cad-cae-copilot copilot-instructions.md
Copilot instructions for armpro24-blip/cad-cae-copilot, covering github copilot — aieng workspace and essentials.
phone-mcp CLAUDE.md
Claude Code instructions for premex-ab/phone-mcp, covering claude.md, project overview, build commands, architecture and module layout.
zmk-config AGENTS.md
AGENTS.md instructions for urob/zmk-config, covering customization guide, ground rules, how the multi-board layout works, adding a new board and where to change what.