cf-mcp
Read-only Cloudflare MCP server for SOC investigation, exposing Cloudflare services like WAF, bot management, analytics, and Zero Trust for querying and discovery.
README
cf-mcp
Read-only Cloudflare MCP server for SOC investigation agents (companion to
sumologic-mcp, flare-mcp, mcp-virustotal).
Exposes:
- Discovery — accounts, zones
- Rulesets & WAF — custom rules, rate-limit rules, managed rulesets
- Bot Management — config, Super Bot Fight Mode
- Analytics (GraphQL) — firewall events, HTTP requests, bot events, baseline traffic diff
- Cloudflare One / Zero Trust — Access apps & policies, Gateway rules, WARP devices, IdPs
- Logpush — job metadata
- Helpers — dashboard URL builder, wirefilter validator
v1 is strictly read-only. Every non-GET HTTP request is refused at the
client layer when CF_READ_ONLY=true (default).
Install
uv sync
Configure
Store your Cloudflare API token in the OS credential store (Windows Credential Manager / macOS Keychain / Linux Secret Service):
uv run cf-mcp-setup
Alternatively, set the CF_API_TOKEN environment variable in your MCP client
config — useful on headless Linux hosts where no keyring backend is available.
Required token scopes
Create an API token at https://dash.cloudflare.com/profile/api-tokens with at least these read permissions:
- Zone Read
- Zone WAF Read
- Account Rulesets Read
- Account Settings Read
- Bot Management Read
- Analytics Read (Account + Zone)
- Logs Read
- Access: Apps Read, Access: Policies Read
- Zero Trust: Gateway Read
Optional env vars
CF_ACCOUNT_ID— default account ID for account-scoped tools.CF_ZONE_ALLOWLIST— comma-separated zone names; zone-scoped tools refuse zones not in the list, even if the token has broader access.CF_READ_ONLY— defaults totrue; setfalseonly if a future v1.5 ships mutating tools and you've reviewed them.
MCP client config
Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"cloudflare": {
"command": "uv",
"args": ["run", "--project", "/path/to/cf-mcp", "cf-mcp"],
"env": {
"CF_ACCOUNT_ID": "<your-account-id>",
"CF_ZONE_ALLOWLIST": "example.com,example.net"
}
}
}
}
Development
uv sync
uv run ruff check src tests
uv run mypy --strict src
uv run pytest tests/unit
CF_LIVE_TEST=true uv run pytest tests/integration
Architecture notes
- One async httpx client per process, shared across all tool calls.
- No retry on 429 — the agent decides;
retry_after_sis surfaced in the error envelope. Retry 502/503/504 with exponential backoff + jitter, max 3 attempts, ~10s total budget. - No internal rate-limit accountant — Cloudflare's edge and the agent are the only governors. REST quota is 1200/5min; GraphQL is a separate 300/5min.
- No cache in v1 — the
cache_metaenvelope field is reserved for v1.5. - Compact-mode by default — GraphQL tools return only dimensions + counts.
Detail drill-down via
verbose=trueoncf_query_firewall_events_rawandcf_query_http_requests_raw. - Hard ~20K-token response ceiling — exceeding tools return
response_too_largewith a hint, never silent truncation.
Response envelope
{
"data": {...},
"next_cursor": "v1.<base64>",
"cache_meta": {"hit": false, "age_s": 0, "ttl_s": 0},
"api_endpoint_called": "POST /graphql",
"correlation_id": "uuid",
"error": null
}
On error, data is null and error has:
{
"code": "rate_limited|auth|not_found|validation|upstream|response_too_large|read_only_violation|zone_not_allowed",
"http_status": 429,
"cf_errors": [{"code": 10000, "message": "..."}],
"retry_after_s": 30,
"hint": "narrow the time range or reduce limit"
}
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.