x-mcp-skill

# X MCP Skill X MCP skill for Vercel `skills` CLI / Codex / Claude Code / Cursor. This skill provides work procedures based on the official X MCP Servers documentation, covering: - X API MCP: `https://api.x.com/mcp` - X Docs MCP: `https://docs.x.com/mcp` - `xurl mcp` bridge OAuth2 PKCE setup - Grok Build / Cursor / Claude Desktop / VS Code / generic MCP client configuration - Explicit confirmation rules before writing operations with X API ## Install Check available skills: ```bash npx skills add nanameru/x-mcp-skill --list ``` Add to current project: ```bash npx skills add nanameru/x-mcp-skill --skill x-mcp ``` Explicitly add for Codex: ```bash npx skills add nanameru/x-mcp-skill --skill x-mcp --agent codex ``` Add globally: ```bash npx skills add nanameru/x-mcp-skill --skill x-mcp --global ``` Generate a one-time use prompt: ```bash npx skills use nanameru/x-mcp-skill@x-mcp ``` ## Notes - Actual usage of X API MCP requires `CLIENT_ID` and `CLIENT_SECRET` from the X Developer App. - In the X Developer Console's User authentication settings, typically select `Web App, Automated App or Bot` and register `http://localhost:8080/callback` as the Callback URI / Redirect URL exactly. - If the OAuth authorization screen displays "Cannot access the app," ensure OAuth 2.0 is enabled, the callback URL matches exactly, App permissions and required scopes are consistent, and OAuth 2.0 Client ID / Client Secret are being used. - Treat `CLIENT_SECRET`, Bearer token, OAuth token, and contents under `~/.xurl` as secrets. - Posting, article publishing, bookmarking, following, DM, and other write/account operations require explicit user confirmation before execution. ## Sources - https://docs.x.com/tools/mcp - https://docs.x.com/skill.md - https://docs.x.com/llms.txt