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
source-command-pr-shepherd
Monitor a PR through to merge - handles CI failures, reviews, and thread resolution
source-command-plan
Restate requirements, assess risks, and create step-by-step implementation plan. WAIT for user CONFIRM before touching any code.
source-command-learn-eval
Extract reusable patterns from the session, self-evaluate quality before saving, and determine the right save location (Global vs Project).
source-command-instinct-status
Show learned instincts (project + global) with confidence
source-command-instinct-import
Import instincts from file or URL into project/global scope
source-command-instinct-export
Export instincts from project/global scope to a file
source-command-evolve
Analyze instincts and suggest or generate evolved structures
source-command-e2e
Generate and run end-to-end tests with Playwright. Creates test journeys, runs tests, captures screenshots/videos/traces, and uploads artifacts.
source-command-docs
Look up current documentation for a library or topic via Context7.
source-command-devfleet
Orchestrate parallel Codex agents via Codex DevFleet — plan projects from natural language, dispatch agents in isolated worktrees, monitor progress, and read structured reports.
source-command-claw
Start NanoClaw v2 — ECC's persistent, zero-dependency REPL with model routing, skill hot-load, branching, compaction, export, and metrics.
source-command-aside
Answer a quick side question without interrupting or losing context from the current task. Resume work automatically after answering.
sora
Use when the user asks to generate, remix, poll, list, download, or delete Sora videos via OpenAI’s video API using the bundled CLI (`scripts/sora.py`), including requests like “generate AI video,” “Sora,” “video remix,” “download video/thumbnail/spritesheet,” and batch video generation; requires `OPENAI_API_KEY` and Sora API access.
social-media-toolchain
Use when working with the user's personal social-media MCP servers (Instagram live, Facebook Messenger groups/pages, WhatsApp proof-of-concept, TikTok public read-only) - which server does what, their auth and setup, the per-platform tool inventory, residential-proxy and account-state guidance, and the hard limits (E2EE, client-version gotchas, ToS and ban risk).
social-find-post
Use when the user wants to find, recall, or locate a specific social-media post, reel, or saved/bookmarked item across Instagram (saved collections, feed, search) or Facebook (feed/profile). Pairs with the social-media-toolchain MCPs.
social-dm-search
Use when the user wants to search, read, or recall direct messages and chats across Instagram DMs, Facebook Messenger group/room/page threads, or WhatsApp chats (proof-of-concept, off by default). Pairs with the social-media-toolchain MCPs.
social-corpus-extraction
Use when building a personal writing/style corpus from a user's OWN messages and posts across Discord, Facebook posts + comments, Facebook Messenger, and WhatsApp — isolating only the owner's authored text, then folding every source into one deduplicated jsonl corpus. Handles the traps that silently break these extractions: the Meta DYI double-encoded-UTF-8 mojibake, the 4-way Facebook post-path drift, WhatsApp-DYI-has-no-message-bodies, and locale-dependent title strings. NOT for scraping other people's content, NOT a live-API social client (see social-media-toolchain), NOT for one specific person's hardcoded identity — every script takes the owner identity by argv/env.
smallest-check-ladder
Use when you need to run "the smallest test that could still catch this change" instead of blindly running the whole suite (or nothing). Provides an invocable L0 to L3 ladder - L0 lint the edited file, L1 impacted tests via deterministic selectors, L2 full suite via longrun, L3 e2e - so a "tests pass" claim is backed by a proportionate check. Never ML "predictive test selection".