BytesAgain Agent API

MCP endpoint (streamable HTTP, works with all MCP 2024-11-05 clients):

# OpenClaw
openclaw mcp set bytesagain '{"url":"https://bytesagain.com/api/mcp","transport":"streamable-http"}'

# Claude Desktop (claude_desktop_config.json)
{
  "mcpServers": {
    "bytesagain": {
      "url": "https://bytesagain.com/api/mcp",
      "transport": "streamable-http"
    }
  }
}

Search 60,000+ AI agent skills by keyword or natural language query. Supports 7 languages: English, Chinese (中文), Japanese (日本語), Korean (한국어), German, French, Portuguese/Spanish. Results are ranked by relevance then download count.

When to use: User wants to find or discover skills for a specific task. Start here before using get_skill.

Returns: Array of skills with: slug, name, description, category, tags, downloads, stars, source, source_url.

curl -X POST https://bytesagain.com/api/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_skills","arguments":{"query":"email automation","limit":5}}}'

{
  "action": "search",
  "query": "email automation",
  "count": 5,
  "results": [
    {
      "slug": "email-automation",
      "name": "Email Automation",
      "description": "Automate email workflows with AI...",
      "category": "email-marketing",
      "tags": ["email", "automation", "workflow"],
      "downloads": 8750,
      "stars": 42,
      "source": "clawhub",
      "source_url": "https://clawhub.ai/skills/email-automation"
    }
  ]
}

Fetch complete details for a single skill by its unique slug. Returns full metadata including version, author, install command, and source URL. Use after search_skills when the user wants more info about a specific skill.

When to use: User has a specific skill slug (from search results) and wants full details or install instructions.

Returns: Full skill object: slug, name, description, category, tags, version, author, downloads, stars, source, source_url, install command. Returns error if slug not found.

curl -X POST https://bytesagain.com/api/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_skill","arguments":{"slug":"chart-generator"}}}'

{
  "action": "get",
  "slug": "chart-generator",
  "skill": {
    "slug": "chart-generator",
    "name": "Chart Generator",
    "description": "Generate charts and visualizations from data...",
    "category": "data",
    "tags": ["chart", "visualization", "data"],
    "version": "2.1.0",
    "author": "bytesagain1",
    "downloads": 14200,
    "stars": 87,
    "source": "clawhub",
    "source_url": "https://clawhub.ai/skills/chart-generator",
    "install": "clawhub install chart-generator"
  }
}

Get the most popular AI agent skills ranked by total download count. Ideal for onboarding users or showing what skills others are using most — without needing a specific search term.

When to use: User wants to discover trending or widely-used skills without a specific topic in mind. Great for first-time users.

Returns: Array of top skills with: slug, name, description, category, downloads, stars, source. Sorted by download count descending.

curl -X POST https://bytesagain.com/api/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"popular_skills","arguments":{"limit":10}}}'

{"action":"popular","count":3,"results":[{"slug":"chart-generator","name":"Chart Generator","downloads":14200},...]}

Search 342 real-world AI use-cases by task or goal. Each use-case links to a curated page of the best skills for that workflow. Use when the user describes what they want to accomplish rather than naming a specific tool.

When to use: User says 'how do I use AI to write weekly reports' or 'what can AI do for my workflow'. Use BEFORE search_skills when the goal is unclear.

Returns: Array of matching use-cases with: slug, title, description, url (bytesagain.com/use-case/slug). Combine with search_skills to find tools for each use-case.

curl -X POST https://bytesagain.com/api/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_use_cases","arguments":{"query":"write weekly reports","limit":5}}}'

{"action":"use_cases","count":3,"results":[{"slug":"weekly-report","title":"Write Weekly Reports","url":"https://bytesagain.com/use-case/weekly-report"},...]}

Search queries are automatically detected and translated to English internally. You can search in any of these languages:

Machine-readable skill index for LLM context injection: