Coding & Dev Tools
Coding agent skills teach AI agents repeatable engineering workflows: reviewing pull requests, writing tests, refactoring safely, and enforcing house style. Install one SKILL.md and your agent applies the same checklist every time, whether you use Claude Code, Cursor, Codex, or another agent.
-
reinforce52 Bundle Project AnalysisProject Analysis Skill(项目系统学习分析)
-
yuzhihe Skill Clarify帮助用户厘清不清楚、不准确或容易混淆的表达,找到合适词语,纠正术语和概念偏差,并在歧义会影响任务结果时对齐含义。 当用户询问“这个词怎么说”“我表达得清楚吗”“帮我理一下概念”“纠正我的表述”“对齐一下”,或明确要求复盘任务中的表达与概念时使用。 当意思已经清楚、只是日常口语或个人风格差异时不要触发;不要在长任务结束时自行追加复盘。
-
yuzhihe Skill Cantonese Vocab Story Generator将不同来源和格式的粤语或中文词语、短语及短句,自适应编排为自然、现代的香港口语助记文章与词汇表。支持繁体中文、粤拼标注、困难条目的补充情境例句、覆盖核验,以及依据历史索引或既有文章排除已使用条目后,从下一 Part 续写新章节。用于“粤语/粵語生词助记文档”“Wordshk 助记故事”“广东话/廣東話串联记忆”“香港口语情境故事”“粤语短句助记”“只用新增词续写”等需求。
-
bizpers11991-code Skill Simulation ValidationValidate numerical simulations against closed-form results so a visualisation can be trusted - choosing integrators, comparing vector fields correctly, testing limits by convergence rather than point values, finite-difference step scans, and the stiff-ODE stability limit that turns RK4 into 1e245.
-
mhollander38 Bundle Check UsageCheck Claude Code's current 5-hour rolling session usage and weekly usage percentages against your subscription limits. Use when the user asks about usage limits, how much of their session/quota they've used, whether they're close to being rate limited, or when to /compact to save context.
-
kitepon Skill UnaiUnai
-
praisondani Skill Save Generated RecipeUnderstand ordinary-language requests to preview recipes, privately save recipes, or add complete meal plans to RecipeBee. Use when the user asks to preview or save a recipe in RecipeBee, manage a RecipeBee recipe image, or import a meal plan into RecipeBee.
-
reinforce52 Skill Topic Storm Builder关键词知识库一站式构建器(STORM 发散 + 深度研究细化 + Obsidian 建库)。当用户想深入了解某个主题/关键词并建立学习型知识库时使用,支持"先发散多角度、再逐类细化、最后建库"。典型触发语:'用知识库建 X'、'用 topic-storm 建 X'、'发散+细化 X'、'把 X 建立成学习知识库'、'研究 X 并建库到 obsidian'、'X 建库'。用户核心诉求是"关键词检索建立知识库"(考研只是举例),知识库要"先总后细"衍生结构。工作流:意图澄清 → STORM 多视角发散 → 知识谱系树 → 逐类深度搜索细化 → 在 E:\obsidian\rein 下按主题名建独立 vault(含 _知识谱系.md、分目录 _总览.md、知识点笔记、_log.md)→ 写后回读校验。
-
reinforce52 Bundle Topic Knowledge Base当用户想深入了解某个主题/关键词并建立学习型知识库时使用,建库采用"先总后细"的知识谱系结构 (_知识谱系.md 全库总图 + 分层目录 + 每类 _总览.md + 知识点笔记 + _log.md 增量日志 + 写后回读校验)。 支持三种信源模式:①web(纯联网搜索,准确系统,~80K tokens)②social(纯社媒多平台采集,真实考生经验,~120K tokens) ③hybrid(默认,双信源融合:Web建知识骨架+社媒注入备考经验/易错点/分歧辨析,~150K tokens,预压缩优化后)。 典型触发语:'研究/学习/了解/深入/科普 X'、'X 是什么'、'建立/创建 X 的知识库'、'把 X 整理成学习资料'、'我想学 X 帮我规划'、 '采集并建库 关键词=xxx'、'多平台采集并建库 xxx'、'媒体汇总/融合建库 xxx'、'建库 关键词=X 模式=web/social/hybrid'。 工作流:先查 E:\obsidian\rein 已有库防重复 → 选择信源模式(默认hybrid)→ 可选社媒采集清洗+预压缩(harvest_clean.py) → web_search 多路检索(social模式跳过)→ 设计知识谱系树(先总后细)→ 在 E:\obsidian\rein\<关键词>\ 按谱系树建目录 → 生成 _知识谱系.md 与各 _总览.md → 双信源融合生成结构化笔记(Web知识点+社媒多平台视角)与 _MOC 总索引 → 写后回读校验 → 引导用户按学习路径逐步学习。 每个主题为独立 vault,路径固定 E:\obsidian\rein\<主题>\,不统一路径。
-
reinforce52 Skill Connect Wechat Qq Channel接入/重扫 OpenClaw 微信(openclaw-weixin)与QQ Bot 通道。Use when 接入QQ/微信、扫码登录、重新扫描。
-
reinforce52 Skill Openclaw Plugin Native RebuildInstall/update an OpenClaw plugin with native deps (better-sqlite3) — rebuild native binaries after install; fixes "Cannot find module better-sqlite3".
-
opspresso Bundle Run DesktopRun and drive the VibeMon Electron app — launch it, put a character on screen, switch between the 2D and 3D engines, and screenshot the window. Use when asked to start the app, take a screenshot of it, check a rendering change in the real app, or interact with its windows.
-
egdev6 Bundle Github Repository BootstrapTrigger: GitHub repository bootstrap, labels, milestones, issue templates, PR templates, Projects v2, project fields, project views. Plan and apply reusable repository bootstrap safely.
-
hayderhekmat Bundle ActorwebappTurn a compatible Apify Actor into a branded web app with a
-
hayderhekmat Bundle Fast Google Maps ScraperHTTP-first Google Maps leads with scoring
-
caoshurong Skill WhichskillsShow which SKILL.md files Claude, Grok, Cursor, Codex and Copilot will actually load from this repo. Use when a skill is not triggering or you are unsure where to put SKILL.md.
-
caoshurong Skill RunbriefWrap noisy test/build commands so the transcript only gets a tail.
-
caoshurong Skill ReviewReview a pull request against the project checklist.
-
riclyme Skill Stata Moderation MarginsGenerates Zelner-style continuous marginal effect plots for moderation hypotheses tested with logit/probit. Produces a multi-panel PNG showing dy/dx of the focal variable evaluated across the full empirical range of each moderator, with rspike 95% CI bars. Use when you say "generate moderation plots", "Zelner plot for H2-H4", or "marginal effects across moderator range". NOT for count models (NBreg) — use coefficient plots instead.
-
hayderhekmat Bundle Google Maps Lead ScraperTurn Google Maps into an automated sales machine
-
hayderhekmat Bundle Amazon Price Rank MonitorAmazon price and visible rank monitor for ASINs and product
-
hayderhekmat Bundle Crm Data Enrichment AgentEnrich CRM contacts and B2B leads with company data
-
hayderhekmat Bundle Local Business Lead FinderDiscover, extract, and enrich local business leads from
-
hayderhekmat Bundle Co Star Real Estate ScraperScrapes publicly accessible pages from CoStar News GB and
-
paradiseengine Skill Paradise ReleasePublish a requested ParadiseEngine release and update the specified consumers after NuGet becomes available.
-
recursivetrail Skill Tony StarkWrites the high-level spec — what to build and tech choices. Use when Nick Fury or the user names Tony Stark, Iron Man, write the spec, high-level design, or tech choices after requirements are approved.
-
recursivetrail Skill Professor XClarifies product requirements by asking questions until nothing is ambiguous. Use when Nick Fury or the user names Professor X, Charles Xavier, refine requirements, clarify the brief, or the need is still vague.
-
lushengguo Skill Tdd TestTDD Test — KCEX Spot C++
-
lushengguo Skill Cpp ProfilingLinux C++ 性能分析专家。使用 perf、火焰图、硬件计数器进行端到端性能诊断。数据驱动,拒绝臆断。Use when diagnosing C++ performance bottlenecks, analyzing CPU hotspots, memory latency, context switches, or when the user asks "why is this slow?"
-
lushengguo Skill Interview Code面试 Coding 陪练
-
lushengguo Skill Spot ProfilingKCEX Spot 做市系统性能分析。结合 perf 火焰图、Prometheus 指标、自定义埋点、jemalloc 内存分析,端到端诊断 800+ symbol 压力测试瓶颈。自动区分冷启动/稳态阶段,分析 per-symbol per-tick 延迟构成。数据驱动,拒绝臆断。Use when 跑 spot benchmark 压力测试、分析 spot 性能瓶颈、做 profiling 优化循环。
-
lushengguo Skill Spot Profiling LoopKCEX Spot 自动 profiling 优化循环。不断 profile → 分析热点 → 尝试优化 → commit 记录 → 再 profile,直到收敛。每次 commit 累积记录尝试了什么、哪些走通了、哪些没走通。Use when 启动自动性能优化循环、跑持续的 profiling 迭代。
-
lushengguo Skill Spot Pre Log Analyze拉取并排查 KCEX spot 预发布(pre)环境日志。用户给出时间区间和要查的问题(如"pre 环境每天凌晨 3:30-8:00 横盘,看是不是下游出了问题"),用 curl 直查 Kibana/ES,结合日志字典定位原因。
-
recursivetrail Skill The WatcherCritiques Marvel SDLC output — requirements, spec, plan, implementation, or ops — before the user gate. Use when Nick Fury, Captain America, or the user names the Watcher, Uatu, review the artifact, or stage review.
-
recursivetrail Skill Captain AmericaLeads implementation — assembles Avenger sub-agents to execute an approved plan in parallel waves and reviews each wave. Use when the user says Captain America, Cap, assemble the Avengers, implement the plan, or build from the architecture plan.
-
naonaoxu Bundle Code Review评审自动化测试代码(断言有效性、fixture 使用、flakiness、步骤复用),输出评审报告。Use when the user asks to review automation test code, 自动化代码评审, 评审测试代码, or review pytest-bdd tests.
Frequently asked questions
What are Coding & Dev Tools agent skills?
Coding agent skills teach AI agents repeatable engineering workflows: reviewing pull requests, writing tests, refactoring safely, and enforcing house style. Install one SKILL.md and your agent applies the same checklist every time, whether you use Claude Code, Cursor, Codex, or another agent.
Which Coding & Dev Tools skills are most installed?
Popular Coding & Dev Tools skills on SkillMD right now include project-analysis, clarify, cantonese-vocab-story-generator. Rankings shift as installs change; sort this page by "Most installs" for the live list.
Do Coding & Dev Tools skills work with Claude Code and Cursor?
Yes. Every skill here ships as a SKILL.md file, an open format that works in Claude Code, Claude.ai, Cursor, Codex, Windsurf, and 60+ other agents. Install one with npx skillmds@latest add <owner>/<name>, or copy the file into your agent's skills directory.