mcp-cloudflare-crunchtools
# MCP Cloudflare CrunchTools
A secure MCP (Model Context Protocol) server for Cloudflare DNS, Transform Rules, Page Rules, and cache management.
## Overview
This MCP server is designed to be:
- **Secure by default** - Threat modeling, input validation, and token protection
- **No third-party services** - Runs locally via stdio, your API token never leaves your machine
- **Cross-platform** - Works on Linux, macOS, and Windows
- **Automatically updated** - GitHub Actions monitor for CVEs and update dependencies
- **Containerized** - Available at `quay.io/crunchtools/mcp-cloudflare` built on [Hummingbird Python](https://quay.io/repository/hummingbird/python) base image
## Naming Convention
| Component | Name |
|-----------|------|
| GitHub repo | [crunchtools/mcp-cloudflare](https://github.com/crunchtools/mcp-cloudflare) |
| Container | `quay.io/crunchtools/mcp-cloudflare` |
| Python package (PyPI) | `mcp-cloudflare-crunchtools` |
| CLI command | `mcp-cloudflare-crunchtools` |
| Module import | `mcp_cloudflare_crunchtools` |
## Why Hummingbird?
The container image is built on the [Hummingbird Python base image](https://quay.io/repository/hummingbird/python) from [Project Hummingbird](https://github.com/hummingbird-project), which provides:
- **Minimal CVE exposure** - Hummingbird images are built with a minimal package set, dramatically reducing the attack surface compared to general-purpose images
- **Regular updates** - Security patches are applied promptly, keeping CVE counts low
- **Optimized for Python** - Pre-configured Python environment with uv package manager for fast, reproducible builds
- **Production-ready** - Designed for production workloads with proper signal handling and non-root user defaults
This means your MCP server runs in a hardened environment with fewer vulnerabilities than typical Python container images
## Features
### Zone Management (2 tools)
- `list_zones` - List all zones accessible by your API token
- `get_zone` - Get zone details by ID or domain name
### DNS Records (5 tools)
- `list_dns_records` - List DNS records with filtering
- `get_dns_record` - Get a single DNS record
- `create_dns_record` - Create A, AAAA, CNAME, MX, TXT, NS, SRV, CAA records
- `update_dns_record` - Update existing records
- `delete_dns_record` - Delete records
### Transform Rules (6 tools)
- `list_request_header_rules` / `set_request_header_rules` - Modify request headers
- `list_response_header_rules` / `set_response_header_rules` - Modify response headers
- `list_url_rewrite_rules` / `set_url_rewrite_rules` - URL path/query rewrites
### Page Rules (4 tools)
- `list_page_rules` - List all page rules
- `create_page_rule` - Create redirects, cache settings, SSL modes
- `update_page_rule` - Modify existing rules
- `delete_page_rule` - Remove rules
### Cache Management (1 tool)
- `purge_cache` - Purge by URL, tag, host, prefix, or everything
## Installation
### With uvx (Recommended)
```bash
uvx mcp-cloudflare-crunchtools
```
### With pip
```bash
pip install mcp-cloudflare-crunchtools
```
### With Container
```bash
podman run -e CLOUDFLARE_API_TOKEN=your_token \
quay.io/crunchtools/mcp-cloudflare
```
## Configuration
### Creating a Cloudflare API Token
1. **Navigate to API Tokens**
- Go to https://dash.cloudflare.com/profile/api-tokens
- Click "Create Token"
- Click "Get started" next to "Create Custom Token"
2. **Configure Token Name**
- Enter: `mcp-cloudflare-crunchtools`
3. **Configure Permissions**
The Permissions section has three dropdowns per row:
- **First dropdown**: Resource type (`Account` or `Zone`)
- **Second dropdown**: Specific permission category
- **Third dropdown**: Access level (`Read` or `Edit`)
Click "+ Add more" to add each permission row. For full management, add:
| Resource | Permission | Access |
|----------|------------|--------|
| Zone | Zone | Read |
| Zone | DNS | Edit |
| Zone | Page Rules | Edit |
| Zone | Transform Rules | Edit |
| Zone | Cache Purge | Purge |
4. **Configure Zone Resources**
- First dropdown: Select "Include"
- Second dropdown: Select "All zones" or "Specific zone"
5. **Configure Client IP Address Filtering (Optional)**
- Click "Use my IP" button to restrict token to your current IP
6. **Create and Copy Token**
- Click "Continue to summary" → "Create Token"
- **IMPORTANT: Copy the token immediately** - it's only shown once!
### Add to Claude Code
```bash
claude mcp add mcp-cloudflare-crunchtools \
--env CLOUDFLARE_API_TOKEN=your_token_here \
-- uvx mcp-cloudflare-crunchtools
```
Or for the container version:
```bash
claude mcp add mcp-cloudflare-crunchtools \
--env CLOUDFLARE_API_TOKEN=your_token_here \
-- podman run -i --rm -e CLOUDFLARE_API_TOKEN quay.io/crunchtools/mcp-cloudflare
```
### Permission Sets by Use Case
#### Read-Only (viewing only)
| Resource | Permission | Access |
|----------|------------|--------|
| Zone | Zone | Read |
| Zone | DNS | Read |
#### DNS Management Only
| Resource | Permission | Access |
|----------|------------|--------|
| Zone | Zone | Read |
| Zone | DNS | Edit |
#### Full Management (all features)
| Resource | Permission | Access |
|----------|------------|--------|
| Zone | Zone | Read |
| Zone | DNS | Edit |
| Zone | Page Rules | Edit |
| Zone | Transform Rules | Edit |
| Zone | Cache Purge | Purge |
## Usage Examples
### List Your Zones
```
User: List my Cloudflare zones
Assistant: [calls list_zones]
```
### Create a DNS Record
```
User: Create an A record for www.example.com pointing to 192.168.1.1
Assistant: [calls create_dns_record with type=A, name=www, content=192.168.1.1]
```
### Add Security Headers
```
User: Add X-Content-Type-Options: nosniff to all responses for zone abc123...
Assistant: [calls set_response_header_rules with appropriate rule]
```
### Purge Cache
```
User: Purge the cache for https://example.com/styles.css
Assistant: [calls purge_cache with files=["https://example.com/styles.css"]]
```
## Security
This server was designed with security as a primary concern. See [SECURITY.md](SECURITY.md) for:
- Threat model and attack vectors
- Defense in depth architecture
- Token handling best practices
- Input validation rules
- Audit logging
### Key Security Features
1. **Token Protection**
- Stored as SecretStr (never accidentally logged)
- Environment variable only (never in files or args)
- Sanitized from all error messages
2. **Input Validation**
- Pydantic models for all inputs
- Allowlist for record types, actions
- Strict format validation for IDs
3. **API Hardening**
- Hardcoded API base URL (prevents SSRF)
- TLS certificate validation
- Request timeouts
- Response size limits
4. **Automated CVE Scanning**
- GitHub Actions scan dependencies weekly
- Automatic PRs for security updates
- Dependabot alerts enabled
## Development
### Setup
```bash
git clone https://github.com/crunchtools/mcp-cloudflare.git
cd mcp-cloudflare
uv sync
```
### Run Tests
```bash
uv run pytest
```
### Lint and Type Check
```bash
uv run ruff check src tests
uv run mypy src
```
### Build Container
```bash
podman build -t mcp-cloudflare .
```
## License
AGPL-3.0-or-later
## Contributing
Contributions welcome! Please read SECURITY.md before submitting security-related changes.
## Links
- [Cloudflare API Documentation](https://developers.cloudflare.com/api/)
- [FastMCP Documentation](https://gofastmcp.com/)
- [MCP Specification](https://modelcontextprotocol.io/)
- [crunchtools.com](https://crunchtools.com)
<!-- mcp-name: io.github.crunchtools/cloudflare -->
TDQS
Scored across 26 tools
Each tool targets a distinct resource+action (DNS, WAF, page rules, header rules, analytics, zones), so an agent can generally pick correctly. However, the many '..._rules_tool' variants (page rules, WAF rules, request/response header rules, URL rewrite rules) share similar verbs and require careful reading of the resource noun to avoid misselection.
Every tool follows the exact same verb_noun_tool convention (create_dns_record_tool, list_waf_rules_tool, purge_cache_tool), with predictable verbs like create/get/list/update/delete/set/purge. No mixing of conventions.
26 tools is slightly heavy but justified by the breadth of the Cloudflare domain (DNS, WAF, page rules, header/URL rewrite rules, analytics, zones, cache). Each tool earns its place with no obvious redundancy, though it sits just above the comfortable range.
DNS, WAF, and page rules all have full CRUD, and analytics/zones have appropriate read operations. Minor gaps exist: header and URL rewrite rules only expose list+set (no single-get or delete), and there is no page-rule get-by-id, but these are workable given 'set' replaces the whole ruleset.