GITHUB AGENT SKILLS / 技能收录
给你的 AI 助手
装上专业技能
我们从 GitHub 持续收录、审核并分类优质的 Agent Skills,按职业与使用场景整理,复制一行命令即可安装到 Claude Code、Cursor 等智能体。
按职业浏览 Agent Skills
为不同岗位挑选的高频技能
按使用场景浏览 Agent Skills
从你要完成的任务出发
按创作者浏览 Agent Skills
按 GitHub Stars 排序的优秀创作者
Microsoft
@microsoftopenclaw
@openclawAnthropic
@anthropicsOpenAI
@openaiThe Apache Software Foundation
@apacheGitHub
@githubAutoGPT
@Significant-GravitasGarry Tan
@garrytan全部 Skills 482556
explore
Explores unfamiliar codebases, diagnoses problems, and provides contextual insights. Analyzes project structure, traces code logic, identifies issues, and generates actionable recommendations. Should be used when the user needs to explore or investigate a codebase — understanding project structure, diagnosing bugs, tracing feature implementation, or assessing code quality. Distinguished from context-learning which focuses on tracing specific code references, this skill focuses on project-wide exploration and diagnosis.
explain-code
Master the art of explaining code clearly: how to distill complex logic into understandable explanations, adjust depth for audience, and make unfamiliar code accessible. Use when user asks to explain code, understand code, document code, or when they need to communicate how code works. Triggers especially when user says "explain this code", "what does this do", "how does this work", "help me understand", or "explain this function".
doublecheck
Three-layer verification pipeline for AI output. Extracts verifiable claims, finds supporting or contradicting sources via web search, runs adversarial review for hallucination patterns, and produces a structured verification report with source links for human review.
debug-test
Master debugging failing tests: when tests fail, how to determine if it's a real bug or test issue, and how to fix test problems without breaking production code. Use when tests fail, test debugging, fixing broken tests, or when tests don't work as expected. Triggers especially when user says "test failed", "debug test", "fix test", "test not working", "why is this test failing".
kubernetes-operator
Use when building a Kubernetes Operator — custom controllers that reconcile CRD state. Triggers on "build an operator", "CRD design", "reconcile loop", "controller-runtime", "kubebuilder", "operator-sdk", "metacontroller", "KOPF", "operator capability levels", or "custom resource". Ships CRD validator, reconcile-loop linter, and OperatorHub capability auditor (all stdlib Python), 4 references on the operator pattern + CRD design + reconcile patterns + tooling landscape, and a /operator-audit slash command. NOT a generic k8s skill — specifically the Operator pattern.
debug-session
Chain-combined skill for systematic debugging workflow. Combines bug reproduction judgment, error analysis, execution tracing, and debugging expertise into a complete session规范. Use when user mentions debugging, fixing bugs, something not working, error messages, or when code isn't behaving as expected. Triggers especially when user says "debug this", "why is this broken", "it's not working", "error message", "something went wrong", or "help me find the bug".
context-learning
Quickly understands unfamiliar code by tracing imports, call chains, and data flows. Produces concise context analysis covering module purpose, dependencies, logic flow, and side effects. Should be used when the user needs to understand code they didn't write — onboarding, investigating bugs, preparing for code review, or figuring out how a feature works. Distinguished from explore which focuses on project-wide investigation, this skill focuses on tracing code references and understanding specific code contexts.
code-comment-writer
Writes precise, reader-oriented code comments explaining "why" not "what". Covers onboarding, refactoring, algorithm explanation, tech debt marking, and API integration scenarios. Should be used when the user mentions adding comments, documenting code, explaining code logic, improving code readability, or needs help understanding code during handoff. Distinguished from refactoring skills which focus on code structure changes.
system-architect
Blueprint(系统架构师)负责整体技术方案设计、架构决策和技术选型。 Should be used when the user mentions designing system architecture, making technology stack decisions, planning microservices decomposition, reviewing architecture designs, or designing high-availability systems. Distinguished from cloud-architect which focuses on cloud-specific infrastructure, and from code-engineer which handles implementation details.
security-auditor
Fortress(安全审计师)专注于代码安全和漏洞检测,保护应用安全。 Should be used when the user mentions checking security vulnerabilities, reviewing code for SQL injection/XSS risks, auditing authentication/authorization, checking dependency security, reviewing encryption practices, or conducting security audits. Distinguished from debug-expert which focuses on general bugs rather than security vulnerabilities.
quality-analyst
Metric(质量分析师)专注于代码质量评估和技术债追踪。 Should be used when the user mentions evaluating code quality, tracking technical debt, generating quality reports, analyzing code coverage, measuring code complexity, or assessing code maintainability. Distinguished from security-auditor which focuses on security vulnerabilities rather than general code quality metrics.
zarr-python
Stores and queries chunked N-D scientific arrays with Zarr-Python 3, including codecs, sharding, S3/GCS storage, and NumPy/Dask/Xarray integration. Use for array layout, bounded I/O, format migration, or scientific metadata preservation.
performance-optimizer
Turbo(性能优化师)专注于应用性能分析和优化,提升系统响应速度。 Should be used when the user mentions analyzing performance bottlenecks, optimizing page load time, reducing API latency, improving React performance, database query optimization, or cache strategies. Distinguished from code-engineer which focuses on writing new functionality rather than optimizing existing code performance.
edge-case-master
Fuzz(边界用例大师)是测试用例生成专家,专注于边界条件、异常场景和压力测试。 Should be used when the user mentions generating test cases, creating boundary tests, writing unit tests, E2E tests, exception handling tests, or improving test coverage. Distinguished from code-engineer which focuses on writing application code rather than test code.
waypoint-bio
Supports work with Outpost Bio's open microbiome foundation models - the Waypoint checkpoints (Waypoint-6m, Waypoint-45m, Waypoint-170m), the Atlas pretraining corpus, the Compass eight-task benchmark, or the `waypoint` CLI from the `waypoint-bio` package. Covers embedding microbiome samples, fine-tuning on taxonomic abundance data, benchmarking a checkpoint on Compass, pretraining a GPT-2 model on taxonomic abundance profiles, and converting MetaPhlAn, Kraken2, QIIME 2, or MGnify abundance tables into waypoint format.
deployment-engineer
Deployer(部署工程师)专注于部署流程自动化、CI/CD 配置和环境管理。 Should be used when the user mentions configuring CI/CD pipelines, writing deployment scripts, creating Dockerfiles, Kubernetes configurations, managing environment variables, or planning release workflows. Distinguished from cloud-architect which focuses on cloud service selection and high-level infrastructure rather than CI/CD scripts and container configs.
debug-expert
Sherlock(调试专家)专注于智能问题定位和修复,像福尔摩斯一样通过细致观察和逻辑推理找到问题根因。 Should be used when the user mentions debugging errors, analyzing crashes, investigating why something is not working, examining error messages/stack traces, or diagnosing unexpected behavior in code. Distinguished from security-auditor which focuses on security vulnerabilities rather than general bugs, and from code-engineer which focuses on writing new code.
venue-templates
Prepares journal manuscripts, conference papers, research posters, and grant documents using venue-specific formatting guidance and bundled LaTeX scaffolds. Supports selecting an official template, checking current page or anonymity rules, adapting academic writing to a venue, or inspecting a submission PDF.