Outline

# Fix list: Outline

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/outline, the October 2026 research run, assessed 8 October 2026. Grade C, 60.3 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 Outline: 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, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 (0). Prices are public as plans by team size, $10, $79 and $249 a month, with nothing per call (10). A new workspace is free for 30 days and becomes read-only until a payment method is added, so no card is needed to start (20). A person creates the workspace and the first key or OAuth approval in a browser. `apiKeys.create` and `/oauth/register` exist but need a signed-in user first (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.

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

Why it scored 67: Graded on the hosted Outline Cloud API with the hosted lines. Status page at status.getoutline.com on Oh Dear with three monitors, Application, Website and Collaboration (20). The history page read No incident history available on 8 October 2026, the seven-day view was clean and the RSS feed was empty, with no uptime figures, so we could not tell a clean record from an unused page and gave 20 of 30. The API documentation says write methods are limited more tightly than reads and gives no numbers. The numbers are in the public source (1,000 requests a minute by default, 25 or 100 a minute on document methods, 1,000 an hour on MCP), which is not a statement about the hosted service (5). Every operation documents 429 with `Retry-After` and three `RateLimit-*` headers, and `documents.update` takes `lastRevision` for a 409 on a stale write. No idempotency keys (12). No SLA found (0). The API and MCP server are generally available (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. Agent ergonomics, 62 out of 100, up to 6.2 more on the total

Why it scored 62: `documents.search` returns a context snippet and ranking with each hit, and list methods take `limit`. No field selection, and hits embed the document object. The MCP server has 19 tools, registered according to the token's scopes (13). `limit` and `offset` paging with `nextPath`, `sort` and `direction`, and structured `filters` on document lists and search (17). Errors return `ok`, `error`, `message` and `status` with codes such as `validation_error` and `rate_limit_exceeded` (14). No idempotency keys. `documents.create` accepts a client-chosen `id`, `documents.update` rejects stale writes with 409 through `lastRevision`, and all 19 MCP tools carry `readOnlyHint` and `idempotentHint`. None carries `destructiveHint` (11). Few required parameters and sensible defaults. No official client libraries (7).

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. Security & auth, 71 out of 100, up to 5.1 more on the total

Why it scored 71: OAuth 2.0 with PKCE (S256), refresh tokens, revocation and dynamic client registration, and API keys that are shown once, revocable, optionally expiring and scoped down to single methods (30). The user guide allows the key in the request payload as `token`. The source also reads `token` from the query string, which the documentation does not mention, so no deduction. A `read` scope, per-method scopes, MCP tools filtered by scope, a workspace switch for MCP, and deletes that go to the trash for 30 days unless `permanent` is set. No confirmation step for writes or deletes (15). Documents and comments are text written by other users. No prompt-injection guidance was found. WebMCP leaves destructive actions out (3). Audit log with actor, IP address and authentication type, kept one year, on every cloud tier and through `events.list` (14). A disclosure policy through GitHub advisories with 26 advisories published and patched in public, and CodeQL in CI. No security.txt, bug bounty, SOC 2 or ISO 27001 found (9). The 2026 scope-bypass advisories are counted under incidents.

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.

## 5. Schema & documentation, 76 out of 100, up to 3.9 more on the total

Why it scored 76: Public OpenAPI 3.0 description at www.getoutline.com/openapi.json and in outline/openapi, 154 operations, and MCP tools typed with zod schemas (25). llms.txt on the vendor site, and changelog and integration pages served as Markdown with a `.md` suffix (10). All 154 operations have a summary and a description. Most are one or two sentences and few say when not to use a method (13). 51 schemas, 60 enums, `maxLength` and `format` constraints and 123 required lists. Scopes on `apiKeys.create` are free strings and request bodies do not close off extra properties (12). 153 example values, a curl and a JavaScript sample in the introduction only, and shared 400, 401, 403, 404 and 429 responses (10). The description is versioned 0.1.0 with no version in the path and no API changelog. The outline/openapi commit log and 19 fields marked deprecated are the record (6).

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 58, provenance 79.

Why it scored 69: The source is public under BSL 1.1, which is not an OSI licence and bars running a competing document service, with each release converting to Apache 2.0 after its change date. The OpenAPI description is BSD-3-Clause. Between closed and open source (22). A privacy policy, a DPA and a security page exist. The policy is undated and states no retention periods, the DPA points to a Data Security and Retention clause that the policy does not contain, and the terms of service cover the website only. Backups are kept at least three months and audit logs one year (14). Deprecated parameters are marked in the OpenAPI description, 19 of them, with no dates and no deprecation policy (5). The DPA lists 13 sub-processors with purpose and country, and the security page places servers in the US on AWS. No change log for the list (17).

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: getoutline.com, registered 2017-11-08 (8 years) (11 of 15)
- Terms of service: read, states 4 of the 7 things a reader expects (7.4 of 10)
- Privacy policy: read, states 2 of the 8 things a reader expects (5.5 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 83 out of 100, up to 1.5 more on the total

Why it scored 83: v1.10.1 was released on 9 September 2026, 29 days before the check, and the main branch had 200 commits since 5 September (30). Five tagged releases since 11 July 2026, and seven dated changelog entries in the same period (20). The 30 newest open issues and pull requests on 8 October 2026 show the maintainer opening and answering items daily, with 83 open in total on a repository of 40,842 stars (22). No vendor entry in the official MCP registry, where a search returned community servers only, and no official client libraries. The OpenAPI description was updated on 7 October 2026 (3). CI, CodeQL and Dependabot are configured in the repository (8).

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.

- 2026-05-07 and 2026-06-15. CVE-2026-43886 (GHSA-7732-6qrg-wjf4, high) let an OAuth client request `read *` and receive wildcard access, fixed in 1.7.0. CVE-2026-54573 (GHSA-5x79-rj4g-qrh8, high) let a scoped API key or token reach other methods by adding a URL fragment, fixed in 1.8.0. Both defeat the scopes an agent's owner relies on. Fixed and published, so 2 points (https://github.com/outline/outline/security/advisories/GHSA-7732-6qrg-wjf4, https://github.com/outline/outline/security/advisories/GHSA-5x79-rj4g-qrh8).
- 2026-06-06 and 2026-07-27. Two MCP advisories, `list_documents` returning metadata of private documents on an exact slug match (GHSA-pp65-6cc2-4mx9, medium) and viewers publishing into restricted collections through MCP (GHSA-c43v-wwv4-9mcc, medium), both fixed in 1.8.0. Fixed and published, 1 point (https://github.com/outline/outline/security/advisories/GHSA-pp65-6cc2-4mx9, https://github.com/outline/outline/security/advisories/GHSA-c43v-wwv4-9mcc).
- 2025-11-16 to 2026-07-30. Eighteen further advisories in twelve months, among them a critical rate-limit bypass allowing brute force of the email login code (CVE-2026-33640, fixed in 1.6.0) and cross-workspace access to private documents through share creation (CVE-2026-41649, fixed in 1.7.0). All fixed and published by the vendor, so 2 points (https://github.com/outline/outline/security/advisories).

## 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 On-Premises price table on the pricing page is drawn in the browser, so self-hosted editions and prices were not read
- unchecked: the rate limits applied on the hosted service. The numbers quoted come from the public source and its defaults
- unchecked: the MCP tool list as served by a live workspace, which needs a signed-in session. The 19 tools were counted on the main branch
- The status history page shows no incidents at all and no uptime figures. Whether that is a clean record or an unused page could not be established
- The only terms document found reads as website terms, yet the DPA treats the Terms of Service as the customer agreement. `provenance.terms` points at it for want of another
- The lead called Outline open source. The repository is under BSL 1.1, which its own text says is not an open source licence
- Whether the hosted service runs the tagged release or the main branch was not established. The pricing page calls it always up to date
- No bug bounty, SOC 2 or ISO 27001 statement was found on the pages read. One may exist under a confidentiality agreement

## Weaknesses

- No rate limit numbers in the API documentation. The limits we found are in the source code, not stated for the hosted service
- Two flaws that let a restricted key or token exceed its scopes were published in May and June 2026, both fixed (GHSA-7732-6qrg-wjf4, GHSA-5x79-rj4g-qrh8)
- No official client libraries. The vendor points to the OpenAPI description for generating clients
- No idempotency keys, and no MCP tool carries a destructive hint or a confirmation step
- The published terms of service read as website terms, and the privacy policy states no retention periods
- No security.txt, bug bounty or SOC 2 or ISO 27001 report was found on the vendor's pages

## 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.

- Send every call as POST with a JSON body to `https://app.getoutline.com/api/<method>`, for example `documents.search`, with a Bearer key in the header
- Ask for a key scoped to the methods the task needs, such as `documents.info documents.search`. An unscoped key has all of its owner's access
- Pass `lastRevision` on `documents.update` so a stale write is rejected with 409, and use `editMode` `patch` with `findText` to change one passage
- Wait for the seconds in `Retry-After` on a 429. Write methods have tighter limits than reads
- Leave `permanent` unset on `documents.delete`. The default moves the document to the trash, where it can be restored for 30 days

## 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.