# Fix list: Windmill API + MCP From Anchor Terminal's listing at https://www.anchorterminal.com/tools/windmill, the October 2026 research run, assessed 1 October 2026. Grade C, 56.1 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 Windmill API + MCP: 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. Reliability, 40 out of 100, up to 12 more on the total Why it scored 40: status.windmill.dev redirects to an UptimeRobot page (20). The page and its JSON endpoint didn't load for our reader (one errored, one is blocked by robots.txt), so we couldn't read the last 90 days (5, and that's our limit, not a finding against Windmill). No API request limit or 429 guidance published. Concurrency limits exist but you set them per script (0 and 0). Enterprise support promises a 3-hour response, but no uptime SLA document was found (5). API and MCP server are generally available. Only AI sessions are marked beta (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. ## 2. Payments & pricing, 35 out of 100, up to 8.1 more on the total Why it scored 35: No x402, MPP or L402 (0). Seat and worker prices are public ($20 a developer, $10 an operator, $50 a standard worker a month) but there's no per-execution price (15). Cloud free workspaces and the self-hosted Community Edition both run unlimited executions without a card (20). A person signs up or deploys an instance (0). 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. ## 3. Security & auth, 60 out of 100, up to 7 more on the total Why it scored 60: Tokens take scopes down to a path (`jobs:run:scripts:u/admin/my_script`), carry an expiry and can be revoked, and MCP signs in with OAuth where the user picks the scope (30). Less 10 because the documented default MCP URL carries the token as `?token=`, which superadmins can switch off (20). Read-only scopes and folder filters limit what the agent sees, with no confirmation step before destructive tools (14). The MCP docs explain that identity read from request headers can't be forged by prompt injection, but say nothing about untrusted script output (8). Audit logs on Enterprise per the pricing page, and job logs for every run on every edition (12). No SECURITY.md and no security.txt found. Four repository advisories were published, but the critical SQL injection (CVE-2026-23696) reached the public through NVD and VulnCheck rather than a Windmill advisory. No SOC 2 report found (6). 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, 73 out of 100, up to 4.4 more on the total Why it scored 73: 42 endpoint tools plus one tool per exposed script and flow, so 5, plus 10 back because tokens can be limited to folders or favourites and to one workspace (15). Page and per_page on list endpoints, with filters on jobs (18). Errors come back as text messages, mostly undocumented in the spec (10). Annotations are set from the HTTP method, `GET` read-only and `DELETE` destructive, and every `POST` is marked destructive whether it is or not. No idempotency keys (15). Few required parameters, and Apache-2.0 clients for TypeScript, Python, Go and Rust (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, 79 out of 100, up to 3.4 more on the total Why it scored 79: OpenAPI 3.0.3 file in the repository with 913 operations, versioned with each release (1.821.0) and Apache-2.0 licensed. Every script and flow gets a JSON Schema from its signature (25). llms.txt per the 30 September check, and `searchDocs` and `readDocsPage` hand the model Markdown docs from inside the MCP server (10). The 42 endpoint tools reuse OpenAPI summaries. The docs tools say when to call them, most others only say what they do (14). Typed parameters with required fields, but flow bodies are OpenFlow objects of nested free-form values (11). 846 response entries are 200s and only about 30 document an error code (6). A release-please CHANGELOG with every release dated, and breaking changes flagged in it, though the API path itself isn't versioned (13). 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. Transparency & trust, 67 out of 100, up to 2.9 more on the total Made of editorial 51, provenance 82. Why it scored 67: AGPL-3.0 core and Apache-2.0 clients and OpenAPI, while the Community Edition images also contain proprietary code. The `LICENSE` file says which is which (25). Terms and privacy pages exist per the 30 September check, and the free edition keeps job details up to 30 days. We didn't read the privacy text this run (10). Breaking changes are flagged in the changelog, but we found no deprecation policy with notice periods (8). The usage-stats code is closed source in the public tree. The source has a setting to turn stats off, and we didn't find a docs page that describes what is sent (8). 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. Provenance checks not met in full (half of this category, computed from checked facts): - Domain age: windmill.dev, registered 2022-01-06 (4 years) (7 of 15) - security.txt: not found (0 of 10) ## 7. Maintenance & community, 87 out of 100, up to 1.1 more on the total Why it scored 87: v1.821.0 released on 1 October 2026 (30). 97 tags in the last 90 days (20). 569 open issues, most of the ten newest unlabelled, including two security-flavoured UI bugs from 26 September and a report of 14 high CVEs in the bundled Go toolchain from 22 September (12). Official clients on npm (`windmill-client`) and PyPI (`wmill`) released with the server (15). CI runs backend tests on Linux and Windows, frontend checks and CLI tests (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. ## Deductions Each comes off the total. A fixed and documented problem counts for less at the next check. - CVE-2026-23696, SQL injection in folder ownership management for any low-privilege user in Windmill 1.276.0 to 1.603.2, CVSS 4.0 score 9.4, exposing JWT secrets and admin identifiers. Fixed in 1.603.3 and published by NVD on 7 April 2026 (https://github.com/advisories/GHSA-34m2-qrpf-6v7q). - GHSA-24fr-44f8-fqwg, published 2 March 2026, high. SUPERADMIN_SECRET could be read publicly through RCE on versions before 1.603.3. CVE-2026-22683, missing authorisation in 1.56.0 to 1.614.0, rated high on 7 April 2026. All fixed, so the deduction is reduced (https://github.com/windmill-labs/windmill/security/advisories, https://github.com/advisories?query=windmill). ## 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. - unchecked: the status page incident history, which didn't render for our reader - unchecked: the privacy policy and data retention text on Cloud - What the Community Edition usage stats contain, since that code isn't public - Whether CVE-2026-47107 (nsjail default permissions, critical, 19 May 2026) affects Windmill's bundled sandbox ## Weaknesses - The default MCP URL puts the token in `?token=` unless a superadmin turns that off - CVE-2026-23696, a critical SQL injection fixed in 1.603.3, had no Windmill advisory - No published API rate limits, 429 guidance or uptime SLA - The OpenAPI file documents errors for only about 30 operations - No SECURITY.md or security.txt ## 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. - Give the agent a token scoped to `jobs:run` on one folder rather than a full user token - Connect over the OAuth gateway or send the token in a header so it stays out of logs - With a multi-workspace token, pass `workspace_id` on every workspace tool - Call `searchDocs` before guessing at a flag or config key, then `readDocsPage` with the returned URL - Poll the job by ID after `runScriptByPath` for long jobs instead of waiting on the call ## What the review panel asked for - advisory for every fix - notice periods - header-only tokens by default - a SECURITY.md file ## 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.