Penpot API + MCP by Penpot (Kaleidos)

# Fix list: Penpot API + MCP

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/penpot, the October 2026 research run, assessed 1 October 2026. Grade E, 43.8 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 Penpot 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. Security & auth, 33 out of 100, up to 11.7 more on the total

Why it scored 33: Personal access tokens can expire after 30 to 180 days or never and can be deleted at any time, but carry no scopes. The hosted MCP URL takes its key in the `userToken` query parameter as the documented setup, and until 2.18.0 (23 September 2026) an MCP key also worked as a full API token, so less 10 (10 of 30). No read-only mode. `execute_code` runs arbitrary JavaScript against the plugin API, which can delete shapes, and the docs only advise starting with read-only operations. The plugin must stay open in the user's tab (5 of 20). Shared files and library content reach the model with no injection guidance (3 of 15). Audit events exist in the codebase, but we found no operator-facing call log (5 of 15). SECURITY.md routes reports through GitHub advisories, and four were published in 2026. No security.txt per the 30 September check, and no bug bounty or certification found (10 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.

## 2. Reliability, 46 out of 100, up to 10.8 more on the total

Why it scored 46: Penpot runs as a cloud service and as self-hosted software, so this is the average of the two rubrics. Cloud (10). No status page per the 30 September check (0), so no readable incident history (5 of 30). No rate limits, 429 guidance or SLA published (0, 0, 0). The RPC API is documented and in use, while the repository's own notes call MCP multi-user mode 'under development and not yet fully integrated' (5 of 10). Self-hosted (82). Official Docker images and install guides (20). Backend, frontend, end-to-end and MCP test workflows run on push and pull request, pass state not visible (20 of 25). 706 open issues, actively labelled with milestones, and several regressions opened on 1 October are marked release blockers for 2.19.0 (15 of 25). CHANGES.md records every release, without an API versioning policy (12 of 15). Version 2.18 (15).

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. Agent ergonomics, 41 out of 100, up to 9.6 more on the total

Why it scored 41: Five MCP tools locally and four on the hosted URL, though `execute_code`'s description is long. Over RPC, `get-file` returns a whole file (18 of 25). We found no documented pagination or field selection for RPC commands (8 of 20). No documented error codes for the API (8 of 20). No readOnlyHint, destructiveHint or other annotations on any MCP tool, and no idempotency keys (2 of 20). No official API client. The Plugin API is typed TypeScript (5 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.

## 4. Payments & pricing, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No machine payment protocol (0). Plan prices are public, Unlimited $7 an editor a month capped at $175 and Enterprise $25 a member, with no per-call price (10 of 20). The cloud Professional plan is free with no card (20). A person signs up and creates the token or MCP key in account settings (0). Self-hosting the community edition is free under MPL-2.0.

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.

## 5. Schema & documentation, 66 out of 100, up to 5.5 more on the total

Why it scored 66: Each instance serves API docs and an OpenAPI description generated from the backend at /api/main/doc, and the MCP tools declare zod schemas (25). No llms.txt. The MCP server's `high_level_overview` and `penpot_api_info` tools hand the model its docs instead (3 of 10). RPC docs are generated with little prose. MCP descriptions are long and tell the model to read the overview before `execute_code` (12 of 20). RPC inputs are typed, but the main MCP tool takes one JavaScript string (9 of 15). A curl example for `get-profile`. No documented error format, and the integration guide says 'we do not have any specific documentation for the webhooks yet' (5 of 15). CHANGES.md per release (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. Transparency & trust, 69 out of 100, up to 2.7 more on the total

Made of editorial 61, provenance 76.

Why it scored 69: MPL-2.0, including the MCP server (30). Privacy policy of 5 August 2025 names Kaleidos Subsidiary S.L., PostHog and Google Analytics, keeps data 'as long as the commercial relationship is maintained', points to a DPA in the terms and says nothing about AI training or data locations (15 of 30). Changes land in CHANGES.md, and the separate penpot-mcp repository was archived with a pointer when it moved, but there's no deprecation policy. The MCP server's move to SDK v2 removed SSE on 23 September with no changelog line we could find (8 of 20). The cloud names some processors without locations, and we didn't check self-hosted telemetry defaults (8 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.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: penpot.app, registered 2020-05-26 (6 years) (11 of 15)
- Status page: not found (0 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 76 out of 100, up to 2.1 more on the total

Why it scored 76: 2.18.1 on 2026-10-01 (30). 2.17.0 (22 July), 2.17.1 (17 August), 2.17.2 (27 August), 2.18.0 (23 September) and 2.18.1 inside 90 days (20). Issues are labelled and given milestones within a day, and an OAuth-for-MCP issue opened on 1 October (18 of 25). Not in the official MCP registry and no official API client (0). MCP tests in CI and dependencies updated on 29 September (8 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.

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- -3: 2026-05-19, three advisories published together, a critical pre-authenticated account takeover through team-invitation tokens (GHSA-4937-35vc-hqjj), an MCP REPL server bound to 0.0.0.0 with an unauthenticated /execute endpoint allowing remote code execution (GHSA-22qr-rp27-j9wm, high) and authenticated SSRF in remote image import (GHSA-35g2-w7f6-8v9h, high). Fixed and published, so the deduction is reduced (https://github.com/penpot/penpot/security).
- -1: 2026-02-16, arbitrary file read through the create-font-variant RPC endpoint (GHSA-xp3f-g8rq-9px2, high). Fixed and published (https://github.com/penpot/penpot/security).
- -1: 2.18.0 (2026-09-23) fixed MCP keys being usable as full API access tokens, while the documented hosted setup puts that key in a URL query string, and fixed the MCP REPL starting in multi-user mode on the main bind address. Fixed in the changelog with no advisory (https://github.com/penpot/penpot/blob/develop/CHANGES.md).

## 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.

- Whether a status page exists for design.penpot.app; none was found in the 30 September check.
- Telemetry defaults for self-hosted instances, which we didn't check this run.
- Whether the SSE removal of 23 September will be called out in the 2.19.0 changelog.
- API rate limits on the cloud instance.

## Weaknesses

- Personal access tokens have no scopes, and the hosted MCP key goes in a URL query string
- No annotations on MCP tools and no read-only mode
- Four advisories in 2026, including MCP REPL remote code execution
- The MCP server needs the Penpot plugin open in a browser tab, so it can't run headless
- No status page, published rate limits or webhook documentation

## 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.

- Call `get-profile` first to check the token, then `get-teams`, `get-projects` and `get-file` to walk down
- Ask for JSON with `Accept: application/json`, since some commands default to Transit
- Call `high_level_overview` and `penpot_api_info` before `execute_code`. They tell the model what the plugin API can do
- Keep the Penpot tab in the foreground. A backgrounded or frozen tab stalls MCP calls
- Give tokens an expiry. They carry full account access

## What the review panel asked for

- Headless MCP mode
- Document the webhooks
- Document errors and pagination

## 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.