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
to-tickets
Split an agreed plan or spec into actionable tickets with acceptance criteria and blocking dependencies. Draft or publish as requested.
to-spec
Turn an existing discussion into an actionable spec. Draft locally or publish to the configured tracker as requested.
tdd
Develop a behavior test-first through red, green, and refactoring. Use when the user requests TDD or a test-first implementation.
setup-matt-pocock-skills
Configure repository issue-tracker, triage-label, and domain-document conventions when setup or reconfiguration is requested.
runpodctl
Manage Runpod pods, serverless endpoints, templates, volumes, and transfers with runpodctl.
resolving-merge-conflicts
Resolve an in-progress Git merge or rebase by preserving both sides' intent and verifying the result.
research
Investigate a question against primary sources and save cited findings as a repository Markdown file.
prototype
Build a runnable prototype to answer a logic, state, or UI design question. Keep production integration outside the task unless requested.
modern-cpp
Guides C++ code toward modern idioms (C++20/23/26). Use when writing new C++ code, modernizing legacy patterns, or working on security-critical C++. Replaces raw pointers with smart pointers, SFINAE with concepts, printf with std::print, error codes with std::expected.
improve-codebase-architecture
Survey a codebase for architecture improvements and present a visual report. Explore a selected candidate when requested.
implement
Implement an existing spec or tickets, verify the requested behavior, and report the result.
grill-with-docs
Interview the user to resolve an unclear plan or design and record agreed domain terms and architectural decisions.
flash
Develop and deploy AI workloads with the runpod-flash SDK and CLI on Runpod serverless GPUs or CPUs.
domain-modeling
Clarify domain terminology and record agreed glossary entries or architectural decisions. Use when changing the domain model.
diagnosing-bugs
Diagnose code failures, intermittent bugs, and performance regressions. For NodeTool workflow-run failures, use nodetool-troubleshooter.
companion-clis
Use Hugging Face, GitHub, Docker, and AWS CLIs for Runpod model downloads, images, releases, and volume transfers.
codebase-design
Design module interfaces and test boundaries, or compare ways to consolidate behavior behind a smaller interface.
code-review
Review a diff, branch, or PR for correctness, repository standards, and spec coverage. Report evidence-backed findings.