pen.dev

# Fix list: pen.dev

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/pen-dev, the October 2026 research run, assessed 8 October 2026. Grade D, 47.6 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 pen.dev: 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, 33 out of 100, up to 13.4 more on the total

Why it scored 33: Read with the local-software lines, because the MCP server and the CLI run on the owner's machine. `@pen.dev/cli` installs from npm with Node.js 22.19 or later stated, and the desktop app has downloads for macOS, Windows and Linux (20). The source is closed, so no public CI or test suite can be seen (0 of 25). The public tracker highagency/pen-desktop-releases has 14 issues opened between 28 July and 6 October 2026, none closed or answered, one of them a report that headless `pen interactive` saved an empty document over a 1.4 MB file with exit status 0. These are users' reports that we did not reproduce (5 of 25). No changelog, empty release notes, and the format page reserves breaking changes (0 of 15). The desktop app is at 1.2.16 and the CLI at 0.3.11 (8 of 15). Sign-in and plan limits also depend on api.pen.dev, which has no status page.

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. Security & auth, 33 out of 100, up to 11.7 more on the total

Why it scored 33: The CLI uses a stored session token or an organisation-scoped CLI key that the npm README says can be created and revoked in Developer Keys. No scopes, expiry or rotation are documented (18 of 30). The MCP server has no read-only mode, and `execute` is annotated destructive. The integrated agent asks permission for file and command actions, with a 'Dangerously skip permissions' switch, and an owner enables each external client by hand (6 of 20). The `browser` tool and web import bring live page content to the model, and we found no prompt-injection guidance (2 of 15). The Agent tab lists tool activity, `--usage` writes token use to JSON and Pro adds usage tracking, with no audit log documented (5 of 15). No security.txt, disclosure policy, bug bounty or certification was found. The privacy policy says stored data is encrypted in transit and at rest (2 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.

## 3. Payments & pricing, 38 out of 100, up to 7.8 more on the total

Why it scored 38: Read with the hosted lines, because the software is free to install but needs a pen.dev account and paid plans lift its limits. No machine payment protocol (0). Plan prices are public, Pro $16 and Ultra $48 a user a month, with no per-call price (10 of 20). The Free plan includes the CLI, MCP access and personal developer keys, and sign-up takes an email address, a name and a username with no card step in the documented flow (20). The npm README documents `pen signup` with flags and `pen login --email --code`, so no browser is needed, but someone must open the verification email and read the code, and organisation keys for CI start at Pro (8 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.

## 4. Schema & documentation, 61 out of 100, up to 6.3 more on the total

Why it scored 61: Each MCP tool we read declares a JSON Schema for its inputs, and the `.pen` format has a TypeScript schema on the docs site (25). No llms.txt. The package ships `SKILL.md` and Markdown reference files that `read_skill` hands to the model (6 of 10). The `execute` description is one line that points to `get_app_state`, with the operations documented only in the skill files, while `browser` states each action and when to use it (12 of 20). `execute` takes a JavaScript snippet as one string. `browser` has enums for action and target (7 of 15). The CLI page has worked examples and a troubleshooting page, with no list of error responses (7 of 15). The npm versions and the document `version` field are the only versioning, with no changelog (4 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.

## 5. Agent ergonomics, 66 out of 100, up to 5.5 more on the total

Why it scored 66: Four standard tools with short definitions and two conditional ones, though the model must load skill files before `execute` is usable (22 of 25). Output size is in the agent's hands through `Get` callbacks with `skipChildren` and `Print`, and screenshots of chosen nodes, with no documented paging (12 of 20). A failed `execute` returns an `editId` so the agent patches the snippet with `edits`, and `--verbose-mcp` prints full tool errors, but the docs warn that an export failure can exit with status 0 (10 of 20). All tools carry readOnlyHint, destructiveHint, idempotentHint and openWorldHint annotations in the 0.3.11 package, and `--in` with a separate `--out` leaves the source untouched. No idempotency keys (14 of 20). Few required parameters and a one-line headless command. No SDK in any language (8 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.

## 6. Transparency & trust, 53 out of 100, up to 4.1 more on the total

Made of editorial 54, provenance 51.

Why it scored 53: Closed source with clear terms, an EULA and a proprietary licence file in the npm package (15 of 30). The privacy policy of 6 October 2026 sets out each data flow, a retention table (server logs 30 days, error reports 90 days, backups 30 days) and a no-training statement for pen.dev Pro models, and it agrees with the terms, the EULA and the sub-processor list. A DPA is on request only (24 of 30). The old `@pencil.dev/cli` package carries a deprecation notice pointing to the new one. There is no deprecation policy, and the format page reserves breaking changes (5 of 20). PostHog and Sentry are disclosed with what they exclude. Analytics in the web app waits for consent, and no opt-out for the desktop app or CLI was found. Sub-processors are named without locations (10 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: pen.dev, registered 2025-09-10 (1 year) (3 of 15)
- Terms of service: read, states 5 of the 7 things a reader expects, and has 2 clauses that cost points (4.3 of 10)
- Privacy policy: read, states 6 of the 8 things a reader expects (8.5 of 10)
- Status page: not found (0 of 10)
- Changelog: not found (0 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 59 out of 100, up to 3.6 more on the total

Why it scored 59: `@pen.dev/cli` 0.3.11, desktop v1.2.16 and extension 0.6.75 were all published on 8 October 2026 (30). The CLI had 13 versions between 9 July and 8 October (20). None of the 14 issues in the current tracker has a reply, and one of 25 in the older pencil-desktop-releases tracker was closed. The site links a Discord server, which we did not read (4 of 25). A search of the official MCP registry for 'pencil' and 'pen.dev' returned no entry from the vendor (0). Dependencies in 0.3.11 are current major versions, with no public CI (5 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.

## 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.

- The lead named the product 'pen.dev (Pencil)' at www.pencil.dev. The vendor is High Agency, Inc., the product is now pen.dev, and pencil.dev answers with the pen.dev site.
- unchecked: the Developer Keys screen on app.pen.dev sits behind sign-in, so key scopes, expiry and revocation were read only from the docs and the npm README.
- unchecked: the Discord server, so support responsiveness there is unknown.
- The 14 open issues are users' reports. We reproduced none of them, including a report of 30 September that desktop 1.2.15 could not sign in after the rename.
- Whether the desktop app or CLI has a telemetry opt-out. None was found in the docs or policies.
- Whether a card is asked for at sign-up in the web app. The documented flows mention none.
- The MCP tool annotations were read from the minified 0.3.11 npm bundle, where spawn_agents and browser also carry them. The desktop app's bundled server was not inspected.
- No first release date was established. The earliest dated artefact found is `@pencil.dev/cli` on npm, created 20 March 2026.

## Weaknesses

- No changelog. The three desktop releases on GitHub have empty notes, and the format docs reserve the right to make breaking changes
- 14 issues opened between 28 July and 6 October 2026 in highagency/pen-desktop-releases had no comment or closure on 8 October
- No status page, security.txt, disclosure policy, bug bounty or certification was found
- `execute` takes one JavaScript snippet and carries destructiveHint true, with no read-only mode for an external MCP client
- Closed source under a proprietary licence that forbids reverse engineering, and every surface needs a pen.dev account

## 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 `read_skill()`, then `read_skill({ path: "pen-schema.md" })` and `read_skill({ path: "execute.md" })` before the first `execute`. The tool description alone doesn't document the operations
- Call `get_app_state()` and confirm the active document before editing. The MCP server works on whichever `.pen` file is open in the app
- In headless `pen interactive`, call `save()` before `exit()`, and keep `--in` and `--out` on different paths to preserve the source
- Check that an export file exists. The docs say an export failure can print an error without a nonzero exit status
- In CI set `PEN_CLI_KEY` plus a provider key such as `ANTHROPIC_API_KEY`. Run `pen version`, since `pen --version` is not a flag

## 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.