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
caveman
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested. Integrated into Cavekit: enabled by default for build, inspect, and subagent phases via caveman_mode config. See scripts/bp-config.sh for caveman_mode and caveman_phases.
caveman-internal
Internal token-compression protocol for Cavekit agent artifacts — handoff notes, artifact summaries, context bundles, review notes, status lines. NOT the user-facing /caveman skill; that one compresses assistant replies to the user. This skill compresses machine-to-machine prose so the next agent in the loop reads fewer tokens. Three intensities (lite / full / ultra) with automatic selection based on budget pressure. Used by task-builder, builder, inspector, verifier, convergence-monitor, and the stop-hook status block.
cavekit-writing
How to write Cavekit-quality kits that AI agents can consume effectively. Covers implementation-agnostic cavekit design, testable acceptance criteria, hierarchical structure, cross-referencing, cavekit templates, greenfield and rewrite patterns, cavekit compaction, and gap analysis. Trigger phrases: "write kits", "create kits", "cavekit this out", "define requirements for agents", "how to write kits for AI"
capability-discovery
Detect which MCP servers, Claude Code plugins, and CLI tools are available in the current environment, so kits and build sites can bind to real capabilities instead of imagined ones. Runs via /ck:init --tools-only or `cavekit-tools.cjs discover`. Writes .cavekit/capabilities.json. Trigger phrases: "what do we have available", "what's installed", "detect tools", "can we use X", "setup tools".
brownfield-adoption
Step-by-step process for adopting Cavekit on an existing codebase. Covers the 6-step brownfield process, bootstrap prompt design, spec validation against existing behavior, and the decision between brownfield adoption vs deliberate rewrite. Trigger phrases: "brownfield", "existing codebase", "add Cavekit to existing project", "adopt Cavekit", "layer kits on code", "retrofit kits"
blueprint-writing
How to write Blueprint-quality blueprints that AI agents can consume effectively. Covers implementation-agnostic blueprint design, testable acceptance criteria, hierarchical structure, cross-referencing, blueprint templates, greenfield and rewrite patterns, blueprint compaction, and gap analysis. Trigger phrases: "write blueprints", "create blueprints", "blueprint this out", "define requirements for agents", "how to write blueprints for AI"
autonomous-loop
How Cavekit's autonomous execution loop works — state machine, stop hook, completion sentinel, lock, budgets, iteration cap. Read this skill any time you are about to run /ck:make, debug a stuck session, or write a new command that needs the loop. Trigger phrases: "how does the loop work", "stop hook", "autonomous", "completion sentinel", "why is the session looping".
joxo
Help the person set up Joxo or join a teammate's Joxo project. Use when asked to set up or join Joxo, or when given a Joxo invitation or pairing link.
setup-shell
Install or refresh Agent Guard's plugin-local shell integration. Use when the user explicitly asks to enable or update shell command masking, or when a shell-integration warning directs them to rerun setup-shell.
setup-agent-guard
Diagnose, install, and verify Agent Guard's plugin-local binary, jq and gitleaks dependencies, the active host integration, and live hook protection. Use when Agent Guard reports degraded protection, a SessionStart warning asks for setup, plugin hooks fail or appear bypassed, or a user asks to finish or repair Agent Guard installation.
suede-workflow-skills
Route requests across the full public Suede AI pack to the narrowest specialist and preserve the request verbatim. Use when asked "which Suede skill should I use?", when work crosses lanes, or when only the umbrella is installed. For next-action help: what’s next, recommend the next move, give me the prompt, expand prompt, or make it granular. NOT FOR: one clear lane (use its specialist), one multi-file implementation graph (use suede-graph-flo-xr), or parallel file ownership (use suede-agent-teams).
suede-visibility-grader
Grade one public page A-F across findability, first-screen clarity, CTA pull, proof and trust, AI citation readiness, and design signal, with up to five ranked fixes. Use when asked to "grade this page before promotion" or give a quick readiness check of a landing page, README, docs, or campaign. NOT FOR: deep SEO and schema audits (use suede-seo-audit), AI crawler/citation measurement (use suede-ai-seo), or implementing conversion changes (use suede-site-alchemy).
suede-video
Plan or produce marketing video with scripts, storyboards, shot and asset lists, consent-aware AI or avatar choices, programmatic rendering, edit decisions, pacing, export specs, and video-search metadata. Use when asked to "turn this demo into a short video" or build a repeatable production pipeline. NOT FOR: social calendars (use suede-social), clip-to-guide bridges (use suede-clip-to-guide), paid ad concepts (use suede-ad-creative), still images (use suede-image), or unapproved publishing.
suede-sync-packaging
Prepare a song for sync review: scene fit, supervisor one-sheet, asset checklist, lyric flags, mood tags, AI disclosure, and open clearance questions. Use when asked to "make a sync one-sheet" or package a track for film, TV, ads, games, trailers, or creator campaigns. Drafts pitch materials with open clearance questions labeled. NOT FOR: rights investigations or transfer packages (use suede-rights-passport), folder linting (use suede-release-linter), or rollouts (use suede-campaign-in-a-box).
suede-social
Plan, audit, and draft organic social for founder/brand accounts: platform mix, 30-post audits, calendars, Instagram, LinkedIn/X, short-video scripts, social search, and listening. Use when asked to "audit our Instagram growth" or build an approval-ready content loop. NOT FOR: paid ads (use suede-ads or suede-ad-creative), editorial agents (use suede-newsroom), clip-to-guide work (use suede-clip-to-guide), video rendering (use suede-video), or broad content strategy (use suede-content-strategy).
suede-sms
Design consent-aware SMS/MMS programs: channel fit, welcome, cart recovery, post-purchase and win-back flows, exact copy, segment counts, cadence, quiet hours, platform selection, and measurement. Use when asked to "draft our SMS welcome flow" or audit a text program. Checks current consent, opt-out, and registration requirements. NOT FOR: email (use suede-emails), capture UX (use suede-site-alchemy), legal certification, or sending, scheduling, importing, or registering anything.
suede-site-alchemy
Turn landing, launch, campaign, and product pages into measured conversion paths: funnel stage, friction, hero, proof, CTAs, pricing, mobile QA, and pre-registered experiments. Use when asked to "fix this page that gets no signups" or improve demos and sales from existing traffic. Implements and verifies changes within scope. NOT FOR: SEO scores (use suede-seo-audit), AI citations (use suede-ai-seo), promotion grades (use suede-visibility-grader), or copy alone (use suede-copy).
suede-ship-copy
Produce one high-stakes public draft with parallel research, agent-claim audits, message ownership, adversarial review, publish readiness, and an evidence record. Use when asked to "research and draft this launch page" or rigorously develop one landing page, email, post, README, ad, or store listing. Preserves requester-supplied facts and writes drafts only. NOT FOR: routine or bulk copy (use suede-copy), code changes (use suede-graph-flo-xr), or cleaning existing prose (use suede-deslop).