Agent skill

Wechat Devtools

by WaterTian in WaterTian/wechat-devtools-mcp

“微信开发者工具 MCP —— 小程序构建、预览、调试与自动化测试”

— description from SKILL.md by WaterTian
MITAuto-check passedProductivity & Automation

Install Wechat Devtools

skills CLI
$ npx skills add WaterTian/wechat-devtools-mcp --skill wechat-devtools -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install WaterTian/wechat-devtools-mcp wechat-devtools --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/WaterTian/wechat-devtools-mcp.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/wechat-devtools .claude/skills/wechat-devtools && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
wechat-devtools
GitHub stars
150
Token cost
~3.5k tokens
SKILL.md length
920 words
Files
2 (incl. references)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

  • Works in 2 steps: :安装与配置 → :运行时环境检查
  • SKILL.md covers 前置条件, API 速查表, SOP 标准操作流程 and CDP 日志策略, plus 3 more sections
  • Calls uv, pip and curl

About this skill

Wechat Devtools is a skill in WaterTian/wechat-devtools-mcp (150 stars). Its SKILL.md is about 3.5k tokens, with 1 other file in the folder (references). Licence: MIT.

Workflow steps

2 steps, taken from the step headings in SKILL.md.

  1. :安装与配置
  2. :运行时环境检查

What it can do on your machine

Read from SKILL.md and the folder at commit 31fbf94. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • uv
    • pip
    • curl

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Wechat Devtools loads about 3.5k tokens when it runs, and up to ~8.5k if it reads all its reference files. Until then it costs about 12 tokens; SKILL.md has 920 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~12
When it runs · the whole SKILL.md, loaded when a task matches
~3.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~8.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from WaterTian/wechat-devtools-mcp at commit 31fbf94, republished under its MIT licence (© WaterTian). 920 words, ~3,490 tokens.

Download SKILL.mdSave it as .claude/skills/wechat-devtools/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
wechat-devtools
description
微信开发者工具 MCP —— 小程序构建、预览、调试与自动化测试
version
0.9.18

Wechat DevTools MCP Skill (v0.9.18)

前置条件

Step 0:安装与配置
bash
pip install uv
uv tool install wechat-devtools-mcp --force
json
{
  "mcpServers": {
    "wechat-devtools-mcp": {
      "command": "uvx",
      "args": ["wechat-devtools-mcp"],
      "env": {
        "WECHAT_DEVTOOLS_CLI": "C:\\Program Files (x86)\\Tencent\\微信web开发者工具\\cli.bat",
        "WECHAT_PROJECT_PATH": "D:\\Your\\Project\\Path"
      }
    }
  }
}
  • 注册名用 wechat-devtools-mcp(开发者工具 2.x 把自家 MCP bridge 注册成 wechat-devtools,同名会被覆盖)。macOS CLI 路径 /Applications/wechatwebdevtools.app/Contents/MacOS/cli。各编辑器配置见 README。
  • 必须开启服务端口:设置 → 安全设置 → 服务端口。未开启则所有 CLI 操作报 CLI_TIMEOUT。
Step 1:运行时环境检查

先调 wechat_ide(action='status') 一次确认全部前置条件:

检查项字段失败时
CLI 已安装cli_exists: true配置 WECHAT_DEVTOOLS_CLI
服务端口已开启service_port_enabled: truefalse 必然 CLI_TIMEOUT,去设置里打开;null 是读不到,不等于关闭
项目路径有效project_exists: true配置 WECHAT_PROJECT_PATH,须指向含 project.config.json 的根目录
Node.js 可用node_available: true安装 Node.js
版本一致mcp_version == 本文件 versionuv tool upgrade wechat-devtools-mcp 或同步 skill 副本
已登录is_login → logged_in: truelogin(qr_format='terminal') 扫码
官方内建 MCPofficial_mcp.available见下方分流规则
环境分流规则(IDE 2.x 内建 MCP)

开发者工具 2.x(2026-08-18 起为官方 Stable)在 http://127.0.0.1:<ide_port>/mcp 内建 MCP Server(47 个原子工具)。official_mcp.available: true 且官方 MCP 已接入当前 agent 时:

操作交给谁
开关项目、登录、编译、预览、上传、build npm、点击/输入/滚动、云开发官方工具优先,同一操作不要两边各做一遍
长图拼接截图本 skill。官方 simulator_screenshot 只截视口且压到长边 1280 JPEG
CDP 结构化日志(inspector cdp / navigate)本 skill。能回放采集前的历史消息并过滤噪音
SOP C / D / I / J 任务级流程本 skill 编排,基础步骤可调官方工具

available: false(1.06、IDE 未启动、端口漂移)或官方 MCP 未接入时,本 skill 承接全部能力,不要让用户为基础操作去装官方 MCP。

效率原则
  • IDE 只在会话开始 open 一次;改了代码只需 compile → page_data(自动重连 automator),不要重新 open。
  • 没改代码换页面用 evaluate(fn_source="function(){ wx.reLaunch({url:'/pages/x/index'}); return 'ok' }") → page_data。
  • 连接断开先 start → page_data,不要直接走完整恢复。

API 速查表

完整参数与返回字段见 tool_reference.md。

wechat_ide
action功能关键参数 / 返回
open启动 IDE 并打开项目。cdp_enabled=true(默认)会 kill 已运行的 IDE、带 CDP 端口重启,等小程序 target 就绪后做启动健康检查cdp_port(默认 9222,被占用时换,须与 inspector/navigate/compile 一致);返回 ide_runtime、cdp_ready、project_opened;有 error 时 success:false + startup_errors
login / is_login扫码登录 / 查登录态qr_format;logged_in
close / quit关项目窗口 / 退出 IDE无
status环境诊断service_port_enabled、ide_port、official_mcp、mcp_version
wechat_build
action功能关键参数 / 返回
compile编译并捕获 Error/Warning,成功后自动重连 automator(仅默认 9420)cdp_port;返回 errors、warnings、wxml_errors、npm_warning、automator_verified、fatal_errors
preview生成预览二维码qr_format、qr_output(相对路径相对项目根);返回 qr_stale_warning 表示 bundle 可能没刷新
upload上传到微信后台,生产操作version 必填,desc
build_npm构建 npm。新增/更新依赖后必做,否则运行时报 module ... is not defined无
cache_clean清缓存clean_type(默认 compile;all 慎用)

compile_condition 对 tabBar 页可能被 app 路由守卫覆盖,跳转用 evaluate 更可靠。

wechat_automator

先调 start 开启自动化端口,整个会话一次。

action功能必填 / 返回
start开自动化端口,CLI + TCP + WS 三重验证;窗口未加载完时自动重跑 cli auto(最多 3 轮)返回 verified;false 时按 retry_after_ms 重试
tap / input点击 / 输入selector(input 另需 value)
element_info元素 tagName/text/wxml/size/offset,style_prop 时带 styleselector
set_data热更新页面 data,无需重编译data_json;返回 updated_keys
call_method调页面方法method、args_json?;返回 return_value、path
call_wx / mock_wx调 wx API / Mock 返回值(当前会话有效)method(mock 另需 result_json)
evaluate逻辑层执行 JSfn_source(推荐)或 expression;返回 result、mode
page_stack页面栈返回 depth、pages
page_data当前页 dataexpected_path? 会轮询等页面匹配;返回 path、data、path_mismatch
system_info / storage系统信息 / 本地缓存storage 传 key 取值,不传列 keys

evaluate 用法:

  • fn_source 传完整函数源码(function(){...} 或箭头函数),入参放 args_json(JSON 数组)。多语句、声明、return 都由函数体决定,mode: "function"。
  • expression 只传单个表达式。多语句会退回语句模式(全部执行,mode: "statement"),没有 return 时结果为 null 并附 hint。
  • 例:fn_source="function(){ const p=getCurrentPages(); return p[p.length-1].route }"。
wechat_inspector
action功能关键参数
cdpCDP 采集 WXML 警告、渲染层报错、Runtime 错误。能回放采集前的历史消息duration=10、detail_level、max_logs、cdp_port
consoleautomator 事件采集 console 与 JS 异常。只收连接后的事件duration(排查异常 ≥8s)、log_type、tap_selector

排查「刚才报的错」用 cdp;要与交互严格对齐时间线才用 console。

wechat_screenshot
  • full_page(默认 true)长图拼接,false 只截视口,可配 scroll_top;page_path 不匹配时自动跳转;output_path 留空存到项目 screenshots/。
  • 只在用户要求或需要视觉确认时截图;fixed/absolute 弹窗蒙层可能拍不到,以 page_data 为准。

先看 message 有无 ⚠。以下情况图看着连续但不完整:

返回字段含义应对
is_scroll_view_page: true页面靠 scroll-view 滚动,只截到视口用 evaluate 读数据代替视觉确认
truncated: true超分段上限,底部没拍到full_page=false + scroll_top 分段截
content_gaps: N固定头尾吃光重叠,N 处内容丢失调大 overlap(如 150)重试
detection_confident: false固定头尾识别不可靠结果仅供参考

fixed_header / fixed_footer 是识别到的固定区高度(物理像素)。

wechat_navigate
  • page_path 必填,可带 query;tabBar 页自动走 switchTab,其余 reLaunch(返回 navigation_method)。
  • wait_ms 默认 2000,含网络请求的页面建议 3000。clear_logs=true 过滤跳转前的历史 CDP 日志。
  • 返回 current_page、cdp_logs、navigation_mismatch;带 query 且 check_data=true 时数据大面积为空会给 warning(疑似参数名错)。
  • 前提:start 已调用且项目以 cdp_enabled=true 打开。reLaunch 进入的页面云函数调用可能丢上下文,非 tabBar 页优先 evaluate + wx.navigateTo。
wechat_file
action功能必填 / 返回
project_infoproject_config、app_config、directory、app.js/app.wxss 节选无
list_pagesapp.json 全部页面,含文件完整性返回 pages[{path, complete, missing}]、total
read_page页面四件套源码page_path;返回 files{文件名: 内容}、resolved_base
read_file任意单文件,最多 800 行file_path;返回 content、resolved_path、truncated

路径口径统一:先按 miniprogramRoot 解析再回退项目根,list_pages 的输出可直接喂给 read_page。

云函数与云数据库请用 CloudBase MCP。

SOP 标准操作流程

SOP A:初始化
wechat_ide(action='status')                    # 环境诊断
wechat_ide(action='is_login')                  # 未登录 → login(qr_format='terminal')
wechat_ide(action='open', cdp_enabled=True)    # 9222 被占用时加 cdp_port=9223
  ↳ success=false + startup_errors → 先修复再继续
wechat_automator(action='start')               # verified=false → 按 retry_after_ms 重试,期间可先 compile
wechat_build(action='compile')                 # 建立干净基线,自动重连 automator
wechat_automator(action='page_data')           # 验证连接;AppID undefined → project_path 指到子目录了

project_path 必须是含 project.config.json 的根目录,云开发项目的 miniprogram/ 是子目录。

SOP A-2:改代码后
wechat_build(action='compile') → wechat_automator(action='page_data')

不需要重新 open,也不需要 cache_clean。

页面跳转
场景方式
普通页evaluate(fn_source="function(){ wx.navigateTo({url:'/pages/x/x?id=1'}); return 'ok' }")
tabBar 页wechat_navigate(page_path='pages/x/index')
强制重置同上用 wx.reLaunch
跳转后page_data(expected_path='pages/x/x'),校验 path
SOP B:UI 调试
wechat_file(action='list_pages')                            # 拿有效路径
wechat_navigate(page_path='pages/x/index', wait_ms=3000)    # 跳转 + CDP 日志
wechat_automator(action='page_data')                        # path 必须等于目标页
  ↳ 数据异常 → set_data 热更新验证;元素问题 → element_info;需要看图才 screenshot
SOP C:异常排查(报错 / 白屏 / JS 异常)
wechat_automator(action='page_data')                        # ① 关键字段 null → 数据没加载
wechat_automator(action='evaluate', fn_source='function(){ return wx.cloud.callFunction({name:"x",data:{}}) }')
                                                            # ② 直接调 API 拿完整返回,[object Object] 时必用
wechat_inspector(action='cdp', duration=5)                  # ③ 能回放刚才的错误
wechat_build(action='compile')                              # ④ 看 errors / wxml_errors
SOP D:全页面巡检
wechat_build(action='compile')
wechat_file(action='list_pages')
# 逐页顺序执行(禁止并行):
wechat_navigate(page_path=page, wait_ms=3000)
wechat_automator(action='page_data', expected_path=page)    # path 不匹配 → 标记重定向;字段空 → evaluate 诊断
# 汇总:按 page_data 结果输出报告;只在异常页补截图
SOP E:Mock 集成测试(支付 / 权限 / 网络 / 适配)
mock_wx(method='requestPayment', result_json='{"errMsg":"requestPayment:ok"}')
mock_wx(method='getLocation',    result_json='{"latitude":23.1,"longitude":113.3}')
mock_wx(method='request',        result_json='{"errMsg":"request:fail timeout"}')   # 模拟超时
mock_wx(method='getSystemInfo',  result_json='{"theme":"dark","windowWidth":1024}') # 暗色 / 宽屏适配
tap(selector='.pay-btn') → page_data                                                 # 触发并验证

Mock 仅当前会话有效。拦截请求可用 fn_source="function(){ var o=wx.request; wx.request=function(p){ console.log(p.url); return o.apply(wx,arguments) }; return 'ok' }"。

SOP G:带 query 参数的子页面
wechat_file(action='read_page', page_path='pages/x/x')      # 看 onLoad(options) 的参数名
wechat_navigate(page_path='pages/x/x?id=123', wait_ms=3000)
wechat_automator(action='page_data')                        # 大部分为 null → 参数名错,回到第一步
SOP I:跨页面数据一致性
wechat_file(action='list_pages')
# 逐页:navigate → page_data(expected_path=page),提取公共字段(如 points / level)记入比对表
# 比对同名字段:不一致 → evaluate 直接调 API 对比,检查子页面是否走了独立数据链路
SOP J:小程序 + 管理后台并行比对

管理后台走 Playwright MCP,小程序走本 MCP,两者端口不同可并行提取,比对在主进程串行做。automator 9420 独占,同一时刻只能有一个 agent 操作本 MCP。

Show full SKILL.md (366 more words)Show less

CDP 日志策略

场景参数
快速诊断duration=5, detail_level='concise', max_logs=20
深度排查duration=10, detail_level='full', max_logs=100
页面巡检duration=3, detail_level='concise', max_logs=30
  • concise 只回 errors + warnings;summary.errors > 0 再用 full 拿 source 定位,配合 read_file。
  • cdp 会回放采集前的缓冲区(12 秒前的错误也能拿到),console 只收连接后的事件。
  • 已自动过滤:[system]、WAService.js、WAWebview.js、IDE 外壳页;open 的启动检查还过滤 devtools:// 与 ide:///extensions/。
  • 需要自己排除:devtools:// 来源的 console.assert、SharedArrayBufferIssue、wx.saveFile 即将废弃 类框架预警。
  • ide:///extensions/inject/… 来源的 warning 不是噪音,是框架报的真实问题(无效 app.json 字段、API 废弃、WXSS 选择器不合法)。
  • 计数可能被噪音抬高,最终以 page_data 为准。

page_data 必须校验 data.path 等于导航目标;不一致的常见原因:未登录被拦到登录页、云函数失败 fallback 首页、page_path 拼错、onLoad 条件跳转。传 expected_path 可轮询等待匹配,path_mismatch: true 时用 page_stack 看完整栈。

返回值与恢复

成功 {"success": true, "data": {...}, "message": "..."},失败 {"success": false, "error_code": "...", "message": "...", "hint": "..."}。 error_code 只有 6 种:PARAM_MISSING、CLI_NOT_FOUND、PROJECT_PATH_MISSING、NODE_NOT_FOUND、CLI_TIMEOUT、UNKNOWN_ERROR。连接失败、跳转失败等都归 UNKNOWN_ERROR,看 message 与 hint。

连接断开恢复分两级:

  1. 快速:start → page_data。
  2. 完整:open(cdp_enabled=True) → start → compile → page_data。

故障速查

症状原因解决
CLI_TIMEOUT服务端口未开 / IDE 未运行status 看 service_port_enabled;开端口;open
open 返回 startup_errors小程序启动阶段有致命错误先修代码再 open
start 连续 verified=false冷启动 automator WS 握手未就绪按 retry_after_ms 重试;期间先 compile / build_npm
CLI auto 连续 3 次返回成功但端口未监听项目窗口没加载完或已关闭(cli auto 会假成功;纯 CLI open 后约需 15s)稍等再 start;仍失败 open(cdp_enabled=True) 重启后再 start
任意 automator 动作报 Failed connecting to ws://localhost:9420项目窗口已关闭或自动化未开(IDE 2.x 偶发窗口自关)按返回的 hint:先 start,仍失败 open(cdp_enabled=True) 重开
Failed connecting to ws://localhost:9420 / Connection closedautomator 未启动、断开,或项目窗口已关快速恢复失败再完整恢复
CDP 采集失败 / 采到的全是别的东西未以 cdp_enabled 打开,或 9222 被 Chrome 占用(curl 127.0.0.1:9222/json/version 可确认)open(cdp_port=9223),inspector / navigate / compile 用同一端口
Using AppID: undefined / appid missingproject_path 指向子目录 / 未登录改为含 project.config.json 的根目录;is_login
navigate 后 page_data.path 与目标不一致page_path 拼错、被重定向(未登录 / 参数错 / 云函数失败)、switchTab 未完成list_pages 核对;查登录态与 onLoad 逻辑;增大 wait_ms
evaluate 返回 null 且 mode: "statement"expression 走了语句模式没 return改用 fn_source 并显式 return
元素未找到 / Element is obfuscated不在当前页、selector 错、被遮挡page_stack 确认页面;element_info 验证;换父节点
scroll-view 页长图只有一屏automator 无法捕获 scroll-view 内部滚动返回 is_scroll_view_page: true,改用 evaluate 读数据
长图看着连续但少一截固定头尾吃光重叠看 content_gaps,增大 overlap
截图看不到弹窗 / 拍到错误页面overlay 不在同一渲染层 / 截图前页面被重置以 page_data 为准;传 page_path
compile 成功但 IDE 显示红色 WXML 错误WXML 错误走 IDE 内部通道看 wxml_errors;检查中文引号、未闭合标签
运行时报 @babel/runtime/helpers/xxx is not definednpm 依赖更新后未 build_npmbuild_npm → compile,console(duration≥8, log_type='exception') 验证
工具行为异常 / 参数对不上(IDE 2.x)注册名撞车:官方 bridge 也叫 wechat-devtools注册名改 wechat-devtools-mcp;status 的 mcp_version 可确认调到谁
Windows 上 open 后项目没打开 / CDP 连不上(IDE 2.x)旧版 Windows 分支没有 1.x/2.x 判定升级到最新版;仍失败附 ide_runtime 反馈
ide:///extensions/inject/… 的 warning框架报的真实应用问题当真实告警处理,不要过滤

绝对红线

  • ❌ open 返回 startup_errors 后继续测试
  • ❌ 未确认 logged_in: true 就 preview / upload
  • ❌ 对生产项目 cache_clean(clean_type='all')
  • ❌ 脑补运行状态;同一失败操作重试超过 3 次(应转为诊断根因)
  • ❌ 硬 sleep 等待,用 wait_ms 或 page_data(expected_path) 轮询
  • ❌ SOP 里主动截图,只在用户要求或需要视觉确认时截
  • ❌ 没改代码就 compile
  • ❌ navigate 后不校验 page_data.path
  • ❌ 多个 agent 并行使用 wechat_automator(9420 独占)
  • ❌ 用 miniprogram/ 子目录作 project_path;WXML 属性值里用中文引号(工具无法检测)
  • ✅ 自动化 / 截图前先 start;tap / input 前先 element_info 确认元素
  • ✅ upload 前确认版本号递增、build_npm 已执行
  • ✅ compile 后用 page_data 确认 automator 连接

© WaterTian, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in .agents/skills/wechat-devtools of WaterTian/wechat-devtools-mcp.

  • SKILL.md
  • references/tool_reference.md

Open the folder on GitHubat commit 31fbf94

Compare with similar skills

Wechat Devtools next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Wechat Devtools compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Wechat Devtools this skillWaterTian/wechat-devtools-mcp150—~3.5kAutomated safety check: PassMIT
Wechat To Mdbzd6661/wechat-article-for-ai101—~654Automated safety check: PassNone
Minimal Web Baas DemoTencentCloudBase/CloudBase-AI-Toolkit1.1k1 repos~2.2kAutomated safety check: PassMIT
Wechat Publisherwpsnote/wpsnote-skills179—~2.6kAutomated safety check: PassNone
Md2wechatgeekjourneyx/md2wechat-skill3.7k—~3.8kAutomated safety check: PassCustom licence
Linkdigest Social Link Readersickn33/agentic-awesome-skills47k—~3.4kAutomated safety check: PassMIT

Similar skills

  • Wechat To Md

    bzd6661/wechat-article-for-ai

    Convert WeChat Official Account (微信公众号) articles to clean Markdown files with locally downloaded images.

    101 GitHub stars~654 tokensUpdated 7 mo ago
    Productivity & AutomationAuto-check passed
  • Minimal Web Baas Demo

    TencentCloudBase/CloudBase-AI-Toolkit

    Fast path for a minimal CloudBase Web + database demo (最小前后端 / 最小可用 fullstack / Lovable-like BaaS).

    1.1k GitHub starsUsed in 1 repo~2.2k tokens
    Productivity & AutomationAuto-check passed
  • Wechat Publisher

    wpsnote/wpsnote-skills

    【公众号发布助手】将 WPS 笔记排版并导出为微信公众号 HTML. An agent skill from wpsnote/wpsnote-skills.

    179 GitHub stars~2.6k tokensUpdated 4 mo ago
    Productivity & AutomationAuto-check passed
  • Md2wechat

    geekjourneyx/md2wechat-skill

    Convert Markdown to WeChat Official Account HTML. An agent skill from geekjourneyx/md2wechat-skill.

    3.7k GitHub stars~3.8k tokensUpdated 14 days ago
    Media & CreativeAuto-check passed
  • Linkdigest Social Link Reader

    sickn33/agentic-awesome-skills

    Read one public Xiaohongshu, Douyin, TikTok, YouTube, X or WeChat article link into text an agent can use (transcript, image text, key points) via the LinkDigest API or MCP server.

    47k GitHub stars~3.4k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Burp MCP Vuln Check

    langbyyi/CyberStrikeAI-SRC

    Automate low-impact web vulnerability verification through Burp MCP.

    134 GitHub stars~3.1k tokensUpdated yesterday
    SecurityAuto-check passed

Questions about Wechat Devtools

How do I install Wechat Devtools in Claude Code?

Run `npx skills add WaterTian/wechat-devtools-mcp --skill wechat-devtools -a claude-code`. Or copy the skill folder (.agents/skills/wechat-devtools in WaterTian/wechat-devtools-mcp) into .claude/skills/wechat-devtools in your project. Claude Code loads it when a task matches its description.

How do I install Wechat Devtools in Codex?

Run `npx skills add WaterTian/wechat-devtools-mcp --skill wechat-devtools -a codex`. Or copy the skill folder (.agents/skills/wechat-devtools in WaterTian/wechat-devtools-mcp) into .agents/skills/wechat-devtools in your project. Codex loads it when a task matches its description.

Can I use Wechat Devtools in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add WaterTian/wechat-devtools-mcp --skill wechat-devtools -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/wechat-devtools, .gemini/skills/wechat-devtools, .github/skills/wechat-devtools and .opencode/skills/wechat-devtools in your project.

What does Wechat Devtools need to run?

Going by SKILL.md and its folder, Wechat Devtools needs the command-line tools its instructions call (uv, pip and curl).

Does Wechat Devtools access the network?

SKILL.md names 1 domain. As links in the text: github.com. This is read from the text; nothing was executed.

Is Wechat Devtools safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Wechat Devtools use?

Wechat Devtools is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Wechat Devtools use?

About 3.5k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 5k tokens, read only when the agent opens those files.

What are the alternatives to Wechat Devtools?

Skills that share tags, products or a category with Wechat Devtools: Wechat To Md (bzd6661/wechat-article-for-ai, 101 stars), Minimal Web Baas Demo (TencentCloudBase/CloudBase-AI-Toolkit, 1.1k stars), Wechat Publisher (wpsnote/wpsnote-skills, 179 stars) and Md2wechat (geekjourneyx/md2wechat-skill, 3.7k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Wechat Devtools?

WaterTian (a GitHub user) maintains it in WaterTian/wechat-devtools-mcp, which has 150 GitHub stars. The repository was last updated on September 4, 2026.

Source: WaterTian/wechat-devtools-mcp on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.