Does Skill Doc Generator require an API key?
No. The server itself never calls external AI APIs; all reasoning is done by the AI agent. No API key is needed.
@DemboNauta
基本情報
MCP server that turns any documentation URL into a Claude Code skill file. No API key required — the agent does the reasoning, the server handles fetching and saving
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"skill-doc-generator": {
"command": "npx",
"args": [
"skill-doc-generator"
]
}
}
}Fetches a documentation page and returns its text content and all same-domain links. Call this on the initial URL, then decide which linked pages to fetch next based on the returned links.
Saves a Claude Code skill markdown file to disk. Call this once you have generated the skill content from the documentation. IMPORTANT — 500-line limit: SKILL.md must never exceed 500 lines. If the full skill content would exceed 500 lines, split it into multiple files: keep the frontmatter and a concise overview in SKILL.md (under 500 lines), put detailed sections (API reference, examples, patterns, etc.) in separate .md files passed via extra_files, and add a '## Additional References' section in SKILL.md that links to them with relative paths, e.g. [API Reference](./reference.md).
Fetches key files from a GitHub repository (README, docs, examples) using the GitHub API. More reliable than web scraping for repos with good inline documentation. Set GITHUB_TOKEN env var to avoid rate limits (60 req/h unauthenticated, 5000 with token). Use this instead of fetch_doc_page when the user provides a GitHub URL.
Reads an existing skill file from disk. Call this before updating a skill to retrieve its current content, then fetch updated docs and call save_skill with the merged result.
Returns skill-creator instructions so you can create, improve, or evaluate Claude Code skills. Call with no arguments to get the main SKILL.md. Use the 'section' parameter to retrieve subagent instructions (grader, comparator, analyzer) or the JSON schema reference.
Skill Doc Generator is an MCP server that converts any documentation URL into a Claude Code skill file and bundles the full skill‑creator workflow. It requires no API key and works with any MCP‑compatible IDE (Claude Code, Cursor, Windsurf, etc.).
Install directly from npm with Node.js 18+. Add it to your MCP config—either via claude mcp add --transport stdio --scope user skill-doc-generator npx skill-doc-generator or by adding the JSON entry to your IDE’s MCP config file. Then ask your agent to generate a skill from a documentation URL or GitHub repository. Tools include fetch_doc_page, fetch_github_repo, save_skill, read_skill, and get_skill_creator.
No. The server itself never calls external AI APIs; all reasoning is done by the AI agent. No API key is needed.
Node.js 18+ and an MCP‑compatible agentic IDE (Claude Code, Cursor, Windsurf, or VS Code with Copilot).
Set the `GITHUB_TOKEN` environment variable. The `fetch_github_repo` tool uses the GitHub API, and a token helps avoid rate limits.
By default, skills are saved to `~/.claude/skills/`. You can specify a custom `output_dir` when using `save_skill`.
It connects using stdio transport, as shown in the installation commands.
OrangePro local-first CLI + MCP server for behavior mapping, grounded test generation, and dynamic proof.
Modelglass turns "which model should I use?" into a lookup instead of a guess. Seven tools give your agent live pricing across image, language, video, and audio models. Sourced and versioned, not scraped once and forgott

The tools your AI needs to make your games. Turn prompts into production-ready game assets — animated characters with their weapons attached, seamless textures, terrain, VFX, HUDs — and playable game kits Unity, Godot, U

SSH MCP server for AI agents: remote commands, file transfer, log search and server audits through OpenSSH.

Catch intent conflicts before code conflicts. The open-source coordination protocol for coding agents, built above Git.
コメント