Cloudflare MCP Servers by Cloudflare

# Fix list: Cloudflare MCP Servers

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/cloudflare-mcp, the October 2026 research run, assessed 1 October 2026. Grade BB, 71.1 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Cloudflare MCP Servers: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Payments & pricing, 45 out of 100, up to 6.9 more on the total

Why it scored 45: No x402 or other machine payment on these servers (0). The servers are free and usage bills on Cloudflare's own plans, whose prices we didn't fetch this run (15 of 20). Cloudflare has a free plan with no card per the 30 September check (20). The Documentation server works with no account. Every account server needs a person to sign up and approve OAuth (10 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 2. Reliability, 67 out of 100, up to 6.6 more on the total

Why it scored 67: Scored as hosted servers. cloudflarestatus.com is a Statuspage with component history (20). The incidents feed returned only 21 September to 1 October, 38 incidents, all minor or no-impact, including intermittent API authentication errors on 23 September and Durable Objects errors on 23 and 25 September and 1 October. The servers run on Workers and Durable Objects. The history page is paginated by script and we couldn't read July to mid-September, so 12 of 30. API limits published, 1,200 requests per five minutes per user and 200 a second per IP (15). 429s carry `Retry-After`, `Ratelimit` and `Ratelimit-Policy` headers and block for five minutes. The Code Mode server has honoured `Retry-After` since 28 September, but there's no safe-retry guidance for writes (12 of 15). No SLA covering the MCP servers found (0). The developer docs carry no beta label, though the domain servers are versioned 0.x (8 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 3. Security & auth, 74 out of 100, up to 4.6 more on the total

Why it scored 74: OAuth on every account server, and the Code Mode consent page defaults to a read-only scope template. MCP tokens are pinned to the MCP resource and claim only the scopes Cloudflare granted. API tokens can be scoped per permission and revoked (30). Least privilege comes from the token or the template. There's no server-side confirmation, so a full-access grant lets `execute` call any of about 2,500 endpoints, DELETE included. Issue #485 asks what stops an agent changing production DNS and has no reply (13 of 20). Browser Run returns arbitrary web pages as Markdown and the AI Gateway server returns stored prompts and responses. We found no prompt-injection guidance. Model-written code runs in isolated Dynamic Workers, which contains the code but not the content (4 of 15). Account audit logs record API changes, and outbound requests carry `cloudflare-mcp` or `mcp-server-cloudflare/<server>` User-Agents since 25 August and 1 October (12 of 15). security.txt points to a HackerOne programme and a disclosure policy. Certifications weren't checked this run. Issue #401, reporting a possibly unsanitised path in `sandbox.container.app.ts`, has sat unanswered since 19 June (15 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 4. Agent ergonomics, 77 out of 100, up to 3.7 more on the total

Why it scored 77: Code Mode puts the whole Cloudflare API behind 3 tools that Cloudflare puts at about 1,100 tokens, and each domain server is its own small toolset, Radar's 66 tools aside (23 of 25). Tool results are capped at about 6,000 tokens with `--- TRUNCATED ---` markers that keep JSON valid, `?truncateToolResult=false` lifts the cap, and code can filter before returning (18 of 20). Errors pass through the API's messages. Open bug #468 reports list tools returning `[object Object]` (13 of 20). `execute` is marked destructive and the bindings tools carry full hints, but `search` is marked not read-only although it only reads the spec, and Radar, Observability, Browser Run and Containers tools carry none (10 of 20). `account_id` is resolved automatically for account tokens and single-account users, and Cloudflare publishes API SDKs in several languages (13 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 5. Schema & documentation, 78 out of 100, up to 3.6 more on the total

Why it scored 78: Every tool declares a zod input schema, but two open issues since May and June report optional parameters emitted as `{"not":{}}`, which strict validators reject (22 of 25). llms.txt for the whole developer docs site (10). Bindings tools say 'Use this tool when you need to...', and the README has a section on which server to choose, but Radar's 66 tools weren't read in full (15 of 20). Domain tools are typed, while the recommended Code Mode server's `execute` takes a JavaScript string (10 of 15). The README walks through search and execute with code examples. No error catalogue (9 of 15). Changesets and per-server tags in mcp-server-cloudflare and a public changelog site. The Code Mode repository has no tags or releases (12 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 6. Maintenance & community, 69 out of 100, up to 2.7 more on the total

Why it scored 69: The domain servers were last tagged on 2026-08-11, 51 days before the run date. Code Mode commits ran to 28 September with no tags (20). Five tagged releases since 3 July, on 16, 28 and 30 July and 10 and 11 August (20). 40 open issues, many with no comments, including Workers Observability parse failures from July and August and schema bugs from May and June (8 of 25). com.cloudflare.mcp/mcp in the official registry under a DNS-verified namespace, with 28 remotes (15). CI runs lint, tests and Semgrep, but issue #442 reports undici 5.29.0 with 12 advisories (3 high) in the published tree (6 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 7. Transparency & trust, 90 out of 100, up to 0.9 more on the total

Made of editorial 80, provenance 100.

Why it scored 90: Apache-2.0 for both repositories (30). Privacy policy, DPA and a subprocessor list. Nothing specific to how long the MCP servers keep OAuth grants or tool-call data (22 of 30). The Audit Logs and GraphQL servers carry deprecation notices with a named replacement and 'still works for now', and SSE was retired with a 410 and migration text. No removal dates, and the developer docs still list both deprecated servers (12 of 20). The subprocessor page names 12 third parties with locations plus 16 group companies, last updated 1 October 2025 (16 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Incident history for July to mid-September 2026; the history page is script-paginated and the feed returned only the last 50 incidents.
- Whether the Code Mode commits of 28 September are deployed to mcp.cloudflare.com.
- Cloudflare's current plan prices and certifications, which we didn't fetch this run.
- Whether the open path-handling report on sandbox.container.app.ts (issue #401) was triaged privately.

## Weaknesses

- No server-side confirmation before destructive calls once full access is granted
- No prompt-injection guidance for Browser Run pages or AI Gateway logs
- 40 open issues, several bug reports from May to August with no reply
- Annotations are missing on Radar, Observability, Browser Run and Containers tools
- Developer docs still list the deprecated Audit Logs and GraphQL servers

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Start with `https://docs.mcp.cloudflare.com/mcp` for anything documentation-shaped. It needs no auth
- On mcp.cloudflare.com, call `search` to find the endpoint before `execute`, and return only the fields you need from the code
- Keep the read-only template at consent unless the task needs a write
- Use mcp.cloudflare.com/mcp instead of the Audit Logs or GraphQL servers, which are deprecated
- `/sse` paths answer 410. Use `/mcp`

## What the review panel asked for

- removal dates on deprecations
- tagged Code Mode releases
- confirmation on DELETE
- Browser Run injection guidance

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.