Supports retrieval of the latest Gmail monitor summary via the read_monitor_summary tool, integrating email monitoring into the Claude context.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Claude Rules MCP Server search for skills about logging"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Related MCP server: Skills MCP Server
An MCP (Model Context Protocol) server that reads Claude Code skills and rules from the filesystem and exposes them as tools. Supports HTTP and stdio transports with optional token authentication and Cloudflare Tunnel for remote access.
Serves 484 skills with progressive disclosure (metadata first, full content on demand)
17 global rules accessible individually or concatenated
Project-level rule discovery across multiple repositories
Keyword search across all skills and rules
Monitor summary integration (Gmail and court monitors)
Token authentication for HTTP transport
Dual transport: HTTP for containers/web, stdio for Claude Desktop
Node.js 20+
npm 10+
Docker (for container deployment)
npm install claude-rules-mcpOr run directly without installing:
npx claude-rules-mcp initgit clone https://github.com/jeremykenedy/claude-rules-mcp-server.git
cd claude-rules-mcp-server
npm install
npm run buildRun the interactive setup wizard:
npx claude-rules-mcp initThe wizard walks through transport selection, paths, and token auth. It creates a .env file with your configuration.
Then start the server:
npx claude-rules-mcp runTool | Description |
| Returns the full skills manifest (names and descriptions, not file contents) |
| Lists all skills with optional keyword filter |
| Reads the full SKILL.md for a specific skill by name |
| Lists and reads reference files for a skill |
| Lists all 17 global rule names |
| Reads a specific global rule by name |
| Concatenates all 17 rule files into one response |
| Reads project-level CLAUDE.md and .claude/rules/ |
| Scans for projects with .claude/ directories |
| Keyword search across skill descriptions and rule contents |
| Reads the latest Gmail/court monitor summary |
Variable | Default | Description |
|
|
|
|
| HTTP port |
|
| Path to skills, rules, and manifest |
|
| Path to project directories |
| (none) | Token for HTTP authentication (recommended for remote) |
docker build -t claude-rules-mcp:latest .
docker run -d \
--name claude-rules-mcp \
--restart unless-stopped \
-p 3456:3456 \
-e TRANSPORT=http \
-e PORT=3456 \
-e CLAUDE_DATA_PATH=/data/.claude \
-e CLAUDE_PROJECTS_DIR=/data/sites \
-e MCP_SECRET=your-secret-token \
-v /path/to/.claude:/data/.claude:ro \
-v /path/to/sites:/data/sites:ro \
claude-rules-mcp:latestHealth check:
curl http://localhost:3456/healthAdd to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"claude-rules": {
"command": "npx",
"args": ["-y", "claude-rules-mcp", "run"]
}
}
}For remote via SSH:
{
"mcpServers": {
"claude-rules": {
"command": "ssh",
"args": [
"user@your-server",
"cd /path/to/server && TRANSPORT=stdio node dist/index.js"
]
}
}
}Settings > Customize > Connectors > + > Add custom connector:
Name: Claude Rules
URL: https://your-domain.com/mcp?token=your-secret-token# Health check
curl http://localhost:3456/health
# Test with token
curl -X POST "http://localhost:3456/mcp?token=your-token" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0.0"}}}'This package is open-sourced software licensed under the MIT license.
Read the latest monitor summary (Gmail job/legal alerts + Oregon court case updates). Check this at the start of every conversation to see if there are new job leads, legal updates, or urgent items.
Inputs: no named parameters.
Reads and concatenates ALL 17 global rule files into one response. Use this to load your complete global ruleset into context at once. For a single rule, use rules_get_global_rule instead.
Returns: All 17 rules concatenated with section headers.
Inputs: no named parameters.
Reads a specific global rule file by name from ~/.claude/rules/{name}.md.
Args:
rule_name (string): Rule name without .md (e.g. "php", "laravel", "conventions", "testing") Valid: artisan-commands, banners, commands, conventions, echo-broadcasting, horizon, laravel, license, livewire, package-standards, pest, php, readme, socialite, tailwind, testing, vibe-flow
Returns: Full rule file content.
Inputs: rule_name (required).
Returns the full skills-manifest.json — names and descriptions for all 484 skills plus 17 global rule names. Use this first to discover what is available before fetching individual skills or rules. Returns structured JSON: { total_skills, total_rules, skills[], rules[] }
Inputs: no named parameters.
Reads CLAUDE.md and all .claude/rules/*.md files for a specific project.
Args:
project_name (string): Project folder name under ~/sites (e.g. "laravel-auth-modernized", "greenboard")
Returns: Project CLAUDE.md and all project-level rule files concatenated.
Inputs: project_name (required).
Reads the full SKILL.md for a specific skill by name.
Args:
skill_name (string): Exact skill name (e.g. "vue-expert", "php-pro", "laravel-auth")
Returns: Full SKILL.md content.
Inputs: skill_name (required).
Lists or reads reference files for a specific skill. Many skills have 5-7 deep-dive reference docs in their references/ subdirectory.
Args:
skill_name (string): Exact skill name
filename (string, optional): Specific reference filename to read. If omitted, lists all references.
Returns: List of reference filenames, or full content of a specific reference file.
Inputs: skill_name (required), filename.
Lists all 17 global rule names available in ~/.claude/rules/. Rules: artisan-commands, banners, commands, conventions, echo-broadcasting, horizon, laravel, license, livewire, package-standards, pest, php, readme, socialite, tailwind, testing, vibe-flow. Use rules_get_global_rule to read a specific rule.
Inputs: no named parameters.
Scans the projects directory and lists all projects that have a .claude/ directory.
Returns: List of project names with Claude rules configured.
Inputs: no named parameters.
Lists all 484 skills with names and descriptions, read from the manifest (fast — no filesystem scan). Optionally filter by category or keyword.
Args:
filter (string, optional): Filter by name or description keyword (case-insensitive)
Returns: Formatted skill list with descriptions.
Inputs: filter.
Keyword search across skill descriptions (from manifest) and global rule file contents.
Args:
query (string): Search keyword or phrase
scope ("skills" | "rules" | "all"): Where to search (default: "all")
Returns: Matching skill names/descriptions and rule excerpts with line numbers.
Inputs: query (required), scope.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 11 tools
Each tool targets a clearly distinct operation: reading summaries, getting global rules (single or all), listing rules, skills, projects, getting skill references, and searching. No two tools have overlapping purposes.
All tools follow a consistent verb_noun pattern in snake_case, with a uniform 'rules_' prefix for most tools. The exception 'read_monitor_summary' still uses verb_noun, maintaining overall consistency.
11 tools is well-scoped for the server's domain (reading rules, skills, projects, monitor). Each tool provides a necessary distinct operation without redundancy.
The tool surface covers all read operations for rules, skills, projects, and monitor summaries, plus search and listing. While write operations are absent, the server's apparent purpose is read-only access, making it complete for that use case.
Search, browse & read 340+ Claude Code marketing & SEO skills from any MCP client.
Claude Code / MCP skills for the dev pipeline: discover, spec, design, build, ship, operate.
Find MCP servers, Agent Skills and Claude Code plugins: 84k tools, daily live checks, install lines.
Path-scoped team memories, rules and skills for Claude Code, Cursor, Codex and other MCP clients.
