Rulebook MCP Server

The Rulebook MCP server exposes 13 tools for programmatic task and skill management via the Model Context Protocol (stdio transport, JSON-RPC 2.0).

Setup

bash

rulebook mcp init

Starting the Server

bash

rulebook-mcp

Available Tools

Task Management (7 tools)

Tool	Description	Skill Reference
`rulebook_task_create`	Create a new task with directory structure	See `rulebook-task-create` skill
`rulebook_task_list`	List tasks with status filtering	See `rulebook-task-list` skill
`rulebook_task_show`	Show complete task details	See `rulebook-task-show` skill
`rulebook_task_update`	Update task status	See `rulebook-task-update` skill
`rulebook_task_validate`	Validate task format	See `rulebook-task-validate` skill
`rulebook_task_archive`	Archive completed task	See `rulebook-task-archive` skill
`rulebook_task_delete`	Permanently delete task	See `rulebook-task-delete` skill

Skill Management (6 tools)

Tool	Description	Skill Reference
`rulebook_skill_list`	List available skills by category	See `rulebook-skill-list` skill
`rulebook_skill_show`	Show skill details and content	See `rulebook-skill-show` skill
`rulebook_skill_enable`	Enable a skill in project config	See `rulebook-skill-enable` skill
`rulebook_skill_disable`	Disable a skill	See `rulebook-skill-disable` skill
`rulebook_skill_search`	Search skills by query	See `rulebook-skill-search` skill
`rulebook_skill_validate`	Validate skills configuration	See `rulebook-skill-validate` skill

Quick Examples

typescript

// Task workflow
await mcp.rulebook_task_create({ taskId: "add-auth-system" });
await mcp.rulebook_task_update({ taskId: "add-auth-system", status: "in-progress" });
await mcp.rulebook_task_show({ taskId: "add-auth-system" });
await mcp.rulebook_task_validate({ taskId: "add-auth-system" });
await mcp.rulebook_task_archive({ taskId: "add-auth-system" });

// Skill workflow
await mcp.rulebook_skill_list({ category: "languages" });
await mcp.rulebook_skill_search({ query: "typescript" });
await mcp.rulebook_skill_enable({ skillId: "languages/typescript" });
await mcp.rulebook_skill_validate({});

MCP Configuration

For Cursor (.cursor/mcp.json):

json

{
  "mcpServers": {
    "rulebook": {
      "command": "rulebook-mcp",
      "args": [],
      "env": {}
    }
  }
}

For Claude Code (.claude/mcp.json):

json

{
  "mcpServers": {
    "rulebook": {
      "command": "rulebook-mcp",
      "args": [],
      "env": {}
    }
  }
}

Notes

The server uses stdio transport — stdout is reserved for JSON-RPC messages only
All logs go to stderr (use RULEBOOK_MCP_DEBUG=1 for debug output)
The server auto-discovers the .rulebook config by walking up directories
For detailed input schemas, error handling, and usage of each tool, refer to the individual tool skills listed above

Search AI Tools

rulebook-mcp

Install this agent skill to your Project

SKILL.md