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
issue-restructure
Split one GitHub issue into focused replacements, or merge several into one keeper, migrating substance and cross-linking before closing.
issue-create
Files a new GitHub issue: searches for duplicates, picks focused vs umbrella scope, applies real labels, and previews before posting. Use when the user asks to file, open, or track an issue, and when another workflow needs to record deferred work, a bug found mid-task, or a follow-up as an issue.
issue-close
Closes a GitHub issue as resolved, superseded, or duplicate: reads the body and comments, migrates anything still live, cross-links both directions, and previews before closing. Use when the user asks to close an issue, when work that resolves one has merged, or when grooming finds a duplicate or superseded issue.
groom
Groom the GitHub backlog into theme buckets, umbrella status, a prioritized PR train, stale flags, and untracked gaps.
loop-me
Experimental: find the recurring work worth delegating, and grill you into workflow specs for it.
implement-spec
Build a whole spec's tickets in parallel subagents, onto one integration branch.
handoff
Write a handoff document so another session, host, or person can continue the work.
budget
Check the Claude usage/token budget before dispatching subagents or large fan-outs. Reports session and weekly (all-models + per-model) utilization and reset times, warns when near a cap, and guides right-sizing, deferring, and scheduling a wake-up to resume after a usage-limit reset.
agent-standup
Reconciles and coordinates shared work across multiple coding agents and human operators. Use for cross-agent standups, syncs, claims, handoffs, takeovers, wave planning, ownership or dependency collisions, merge trains, and questions about what is actually in flight. Not for a personal daily recap; use /standup for that.
skill-writer
Writes and reviews shared agent skills plus Claude Code Markdown and Codex TOML subagents for this plugin and bootstrapped projects. Covers choosing how prescriptive to be, the house writing style, whether a skill is user- or model-invoked and how skills call each other, writing the description field that controls when a skill fires, splitting content across reference files, the allowed-tools vs tools frontmatter distinction, and harness-kit's own conventions (shared scripts, dual plugin registration and host-specific script paths). Use when creating a new skill or agent, editing an existing one, writing a pointer line in AGENTS.md or CLAUDE.md, or deciding whether something should be a skill, an agent, a script, or nothing at all.
harness
Audits, bootstraps, updates, and improves a project's agent harness: the skills, agents, hooks, and instructions for Claude Code and Codex. Modes: audit, bootstrap, update, add, migrate, retro, hoist; each reads its own protocol reference.
harness-issue
Files a bug report, idea, or proposed change against the upstream harness repo resolved from the project's config, from any project that uses harness-kit. Two modes: open an Issue, or open a PR with a draft fix. Searches for duplicates first and shows the proposed content before filing. Use when a harness skill, agent, hook, script, or doc misbehaves or has a gap worth reporting upstream; when the user asks to file harness feedback or propose a harness fix; and when a harness retro finds an upstream-worthy item.
harness-builder
Analyze a codebase and recommend an agent harness (agents, skills, hooks).
ask-harness
Find which skill or flow fits your situation, with a map of every harness-kit skill.
wizard
Generates an interactive bash wizard that walks a human through steps only they can perform: it opens each page, says what to click and copy, captures the values, and writes them to .env or GitHub secrets. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Not for steps the agent can perform itself.
ui-review
Audits web UI components for accessibility and UX against WCAG 2.1 AA, reporting each violation as `[CRITICAL|IMPORTANT|MINOR] — file:line` with a specific fix. Use when the user asks about accessibility, a11y, WCAG, screen readers, keyboard navigation, or color contrast, or wants a UX pass over components.
tdd
Test-driven development with a red-green loop, one vertical slice at a time, plus what makes a test worth keeping. Use when building a feature or fixing a bug test-first, when the user mentions TDD or red-green-refactor, when writing integration tests, and when another skill implements a change at an agreed seam.
review-pr
Reviews a pull request with two code-reviewer passes, standards and spec (on your own PR, as the gate before merge: fix the findings and record them), or works through unresolved review feedback on an existing PR: fix, commit, push, and reply in thread. Use when the user asks to review a PR or address review comments, and whenever the open-pr skill reaches its agent review or its outside reviews, or finds an open PR already exists.