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/nightsailer/wechat-clawbot/claude-mdgit clone --depth 1 https://github.com/nightsailer/wechat-clawbotWrote 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/nightsailer/wechat-clawbot/claude-md)<a href="https://agentmods.dev/instructions/nightsailer/wechat-clawbot/claude-md"><img src="https://agentmods.dev/badge/instructions/nightsailer/wechat-clawbot/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 5.1 | $0.00855 | $0.00855 |
| Opus 5 | $0.00428 | $0.00428 |
| Sonnet 5 | $0.00171 | $0.00171 |
| Haiku 4.5 | $0.00085 | $0.00085 |
Grade A, and why
wechat-clawbot 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 6d 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Project: wechat-clawbot — WeChat iLink Bot SDK with multi-user, multi-endpoint gateway for AI backends.
WeChat Bot Constraint: Each WeChat account can create only one Bot, and that Bot is exclusively bound to the creator's WeChat account (1:1). Multiple people cannot share a single Bot. The gateway manages multiple Bots (each from a different WeChat user) routing to multiple endpoints.
Architecture
Two operating modes:
- Channel Mode (
claude_channel/) — single-user MCP bridge connecting one WeChat account to one Claude Code session - Gateway Mode (
gateway/) — multi-Bot, multi-endpoint routing gateway: multiple WeChat Bot accounts (each 1:1 bound to its creator's WeChat account) route messages to multiple upstream AI endpoints (MCP SSE, SDK WebSocket, HTTP Webhook)
Key Directories
src/wechat_clawbot/
api/ ilink HTTP API client (getUpdates, sendMessage, getConfig, sendTyping)
auth/ QR login, account storage, credential management
cdn/ AES-128-ECB CDN upload/download pipeline
claude_channel/ Claude Code MCP Channel bridge (CLI + server)
config/ Pydantic configuration schema
gateway/ Multi-Bot, multi-endpoint routing gateway
channels/ Sub-channel implementations (MCP SSE, SDK WebSocket, HTTP Webhook)
admin.py Admin HTTP API (Starlette, Bearer auth)
app.py Main gateway orchestrator (GatewayApp)
cli.py CLI entry point (clawbot-gateway, 25+ subcommands)
config.py gateway.yaml schema (GatewayConfig, Pydantic)
delivery.py SQLite-backed delivery queue
router.py Message routing engine (active-endpoint, @mention, /command)
session.py User session store with file persistence
endpoint_manager.py Endpoint registry with health tracking
invite.py Invite code system
auth.py Authorization module (allowlist/open/invite-code)
archive.py Message archive sidecar (SQLite)
db.py AsyncSQLiteStore base class
types.py Core dataclasses and enums
media/ Media download, MIME types, SILK transcoding
messaging/ Inbound conversion, send pipeline, slash commands
monitor/ Long-poll monitor loop
sdk/ ClawBotClient library for custom bots (WebSocket)
storage/ State directory, sync buffer persistence
util/ Logger, ID generation, sensitive field redaction
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.
- 6d ago First seen · 72 lines · 855 tokens per session scan A cd55cfc49dbc
wechat-clawbot CLAUDE.md is an instructions file published in the GitHub repository nightsailer/wechat-clawbot (11 stars, last pushed 5mo ago), licensed MIT. It adds 855 tokens to every session, about $0.0043 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
WxJava AGENTS.md
AGENTS.md instructions for binarywang/WxJava, covering wxjava agent 指南, 适用范围与指令优先级, 项目概览, 开发流程 and 代码风格与兼容性.
weclaws AGENTS.md
Instructions for yokingma/weclaws: This file sets repository-wide rules for coding agents in WeClaws.
weclaws CLAUDE.md
Instructions for yokingma/weclaws: Read AGENTS.md first. Then read README.md, docs/manuals/README.md, and the relevant workspace PATTERNS.md / CHANGELOG.md.
LangBot AGENTS.md
AGENTS.md instructions for langbot-app/LangBot, covering agents.md, quick facts, essential commands, where to look and cross-repo sdk work.
weapp-vite CLAUDE.md
Claude Code instructions for weapp-vite/weapp-vite, covering claude.md, project overview, vue support, common commands and development.
PowerX AGENTS.md
AGENTS.md instructions for ArtisanCloud/PowerX, covering powerx development guidelines, active technologies, project structure, commands and qa & observability checklist.