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/jiawei686/wechat-dev-mcp/agents-mdgit clone --depth 1 https://github.com/jiawei686/wechat-dev-mcpWrote 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/jiawei686/wechat-dev-mcp/agents-md)<a href="https://agentmods.dev/instructions/jiawei686/wechat-dev-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/jiawei686/wechat-dev-mcp/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 | $0.01124 | $0.01124 |
| Opus 5 | $0.00562 | $0.00562 |
| Sonnet 5 | $0.00225 | $0.00225 |
| Haiku 4.5 | $0.00112 | $0.00112 |
Grade A, and why
wechat-dev-mcp 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 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.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Agent Workflow Guide for WeChat DevTools MCP
This guide helps AI agents (Claude, Cursor, Windsurf) effectively use the WeChat DevTools MCP server.
The server supports two project types, auto-detected from project.config.json:
compileType: "miniprogram"→ mini-program (pages + DOM)compileType: "game"→ mini-game (canvas, no pages/DOM)
Getting Started
Step 1: Establish Connection
// Use 'launch' to open a project (auto-detects type):
{
"name": "launch",
"arguments": {
"projectPath": "/absolute/path/to/your/project",
"cdpEndpoint": "ws://127.0.0.1:9421/<id>" // optional, mini-game deep debug
}
}
// Or 'connect' to attach to an already running DevTools:
{
"name": "connect",
"arguments": {
"wsEndpoint": "ws://localhost:9420",
"projectPath": "/absolute/path/to/your/project"
}
}
Step 2: Wait for Readiness
After launching:
- Mini-program: use
wait_readyorcheck_healthuntilcompilationStatus: "ready". - Mini-game: readiness is detected via
GameGlobalruntime liveness (NOT pages).check_healthshowscompilationStatus: "game_running"andprojectType: "game".
Step 3: Detect Project Type
Always check projectType from check_health (or get_project_type) before choosing tools:
{ "name": "check_health" }
// Response includes: connected, projectType, engineType, pageReady, cdpConnected, tips
projectType: "program"→ use page/DOM tools.projectType: "game"→ useevaluate,call_wx_method,game_*,game_touch,screenshot.
Tool masking:
ListToolsauto-hides mini-program-only tools (navigate, get_element, tap_element, …) when connected to a mini-game, and hides mini-game-only tools when connected to a mini-program. You won't see irrelevant tools.
For Mini-Programs (projectType: "program")
Use: get_page_data, get_element, tap_element, navigate_to, etc.
For Mini-Games (projectType: "game")
Use: evaluate, call_wx_method, game_get_info, game_touch, game_canvas_screenshot,
game_get_fps, game_webgl_debug, game_get_network_logs, game_export_vconsole, screenshot.
Avoid: get_page_data, get_element, tap_element, navigate_to — these require pages and will be rejected.
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 · 115 lines · 1,124 tokens per session scan A 26c58b385bc1
wechat-dev-mcp AGENTS.md is an instructions file published in the GitHub repository jiawei686/wechat-dev-mcp (17 stars, last pushed 1mo ago), licensed MIT. It adds 1,124 tokens to every session, about $0.0056 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
appwrite AGENTS.md
AGENTS.md instructions for appwrite/appwrite, covering appwrite, commands, stack, layout and libraries.
qiankun AGENTS.md
Instructions for umijs/qiankun, covering qiankun project knowledge base, structure, architecture (big picture), commands and build (vite 8 → package javascript; tsc → declarations).
joint CLAUDE.md
Claude Code instructions for clientIO/joint, covering claude.md, project overview, common commands, install dependencies and build distribution files (runs grunt).
qiankun CLAUDE.md
Instructions for umijs/qiankun: This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
ai AGENTS.md
AGENTS.md instructions for vercel/ai, covering agents.md, project overview, repository structure, key directories and core package dependencies.
Claw3D AGENTS.md
Instructions for iamlukethedev/Claw3D, covering agent instructions, cursor cloud specific instructions, service overview, running the app and lint, typecheck, and tests.