经 AI Skill Hub 精选评估，开源AI技能市场 获评「推荐使用」。这款AI工具在功能完整性、社区活跃度和易用性方面表现出色，AI 评分 7.8 分，适合有一定技术背景的用户使用。
开放式AI代理技能市场平台，支持发现、发布和组合AI工作流技能。基于TypeScript和Next.js开发，提供模块化技能库和工作流编排能力，适合AI开发者、产品经理和自动化工程师构建智能应用。
开源AI技能市场 是一款基于 TypeScript 开发的开源工具，专注于 AI工作流、技能市场、agent代理 等核心功能。作为 GitHub 开源项目，它拥有活跃的社区支持和持续的版本迭代，代码完全透明可审计，支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流，都能提供稳定可靠的解决方案。
开放式AI代理技能市场平台，支持发现、发布和组合AI工作流技能。基于TypeScript和Next.js开发，提供模块化技能库和工作流编排能力，适合AI开发者、产品经理和自动化工程师构建智能应用。
开源AI技能市场 是一款基于 TypeScript 开发的开源工具，专注于 AI工作流、技能市场、agent代理 等核心功能。作为 GitHub 开源项目，它拥有活跃的社区支持和持续的版本迭代，代码完全透明可审计，支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流，都能提供稳定可靠的解决方案。
# 方式一：npm 全局安装 npm install -g openagentskill # 方式二：npx 直接运行（无需安装） npx openagentskill --help # 方式三：项目依赖安装 npm install openagentskill # 方式四：从源码运行 git clone https://github.com/Leon-Drq/openagentskill cd openagentskill npm install npm start
# 命令行使用
openagentskill --help
# 基本用法
openagentskill [options] <input>
# Node.js 代码中使用
const openagentskill = require('openagentskill');
const result = await openagentskill.run(options);
console.log(result);
# openagentskill 配置说明 # 查看配置选项 openagentskill --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量（覆盖配置文件） export OPENAGENTSKILL_CONFIG="/path/to/config.yml"
pnpm install
cp .env.example .env.local
Apply SQL files in scripts/ in order. They create the skills catalog, profiles and points ledger, activity feed, agent feedback loop, aggregate stats view, and RLS policies.
Production writes should go through the Next.js API routes. Public feedback, reviewed skill submissions, and indexer writes use narrow Supabase RPCs guarded by server secrets, while public clients can only read approved skills and aggregate stats directly.
The hourly indexer defaults to high-star, skill-only bulk discovery. It scans GitHub repositories matching agent skill, browser automation, RAG, workflow, and developer-tool signals, excludes MCP repositories, then imports up to INDEXER_RUN_TARGET new approved skills per run. Use GET /api/indexer/logs with automation auth to inspect recent run summaries. Use POST /api/indexer/run with mode: "reviewed" for the slower README/AI-review path.
X posting supports two paths. The free, manual path is /api/x/intent, which generates a scenario-led OpenAgentSkill Update draft and redirects to the official X Web Intent composer; the user must click Post, and no X API credits are used. The main draft includes the skill URL so it can be moved into a reply before posting; /api/x/intent/reply?tweet_id=<posted-tweet-id> can also prefill a manual reply with the same skill URL. The paid API path uses OAuth 2.0 with PKCE: visit /api/x/auth while signed in as the allowed X account to store an encrypted refresh token, then /api/x/post-daily can publish one high-quality unposted skill when called with automation auth. It is intentionally not scheduled in vercel.json.
---
curl -X POST https://www.openagentskill.com/api/agent/feedback \
-H "Content-Type: application/json" \
-d '{
"skill_slug": "browser-use",
"agent_id": "claude-3.5",
"success": true,
"latency_ms": 1200
}'{
"skill_slug": "browser-use",
"total_calls": 12847,
"success_rate": 94.2,
"avg_latency_ms": 850,
"unique_agents": 156
}---
| Variable | Description |
|---|---|
NEXT_PUBLIC_SUPABASE_URL | Supabase project URL |
NEXT_PUBLIC_SUPABASE_ANON_KEY | Supabase anonymous key |
SUPABASE_SECRET_KEY / SUPABASE_SERVICE_ROLE_KEY | Server-only Supabase key for privileged automation/admin routes |
GITHUB_TOKEN | GitHub token for API access |
INDEXER_SECRET | Bearer secret required for indexer/blog routes and reviewed skill-submission RPC writes in production |
INDEXER_RUN_TARGET | Optional. Hourly bulk-import target per run, defaults to 10, max 500 |
INDEXER_MIN_STARS | Optional. Minimum GitHub stars for bulk import, defaults to 500 |
INDEXER_MAX_SEARCH_REQUESTS | Optional. GitHub search requests per bulk run, defaults to 10 without GITHUB_TOKEN |
CRON_SECRET | Bearer secret required for cron-triggered maintenance routes in production |
X_CLIENT_ID | X OAuth 2.0 client ID for authorizing the posting account |
X_CLIENT_SECRET | X OAuth 2.0 client secret for token exchange and refresh |
X_ALLOWED_USERNAME | Optional. X username allowed for OAuth storage, defaults to openagentskill |
The unique feature that sets us apart. Track real agent usage, not just GitHub stars.
| Skill | Stars | Description | Use Case |
|---|---|---|---|
| [Composio](https://github.com/ComposioHQ/composio) | 15K | 250+ app connectors for AI agents | Connect agents to Slack, GitHub, Notion, etc. |
| [Google Workspace CLI](https://github.com/googleworkspace/cli) | 11K | CLI for Google Workspace APIs | Automate Gmail, Docs, Calendar with agents |
创新的AI技能市场生态，提供模块化和可复用的工作流解决方案，社区驱动发展模式有潜力但需要更多生态参与。
AI Skill Hub 为第三方内容聚合平台，本页面信息基于公开数据整理，不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后，再部署至生产环境，并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一，可自由商用、修改、分发，仅需保留版权声明。
AI Skill Hub 点评：开源AI技能市场 的核心功能完整，质量良好。对于AI爱好者来说，这是一个值得纳入个人工具库的选择。建议先在非生产环境试用，再逐步推广。
| 原始名称 | openagentskill |
| 原始描述 | 开源AI工作流：The open marketplace for AI agent skills. Discover, publish, and compose skills 。⭐167 · TypeScript |
| Topics | AI工作流技能市场agent代理开源平台工作流编排 |
| GitHub | https://github.com/Leon-Drq/openagentskill |
| License | MIT |
| 语言 | TypeScript |
收录时间：2026-05-22 · 更新时间：2026-05-22 · License：MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。