# Fix list: Navige From Anchor Terminal's listing at https://www.anchorterminal.com/tools/navige, the October 2026 research run, assessed 17 October 2026. Grade E, 42.8 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 Navige: 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, 28 out of 100, up to 14.4 more on the total Why it scored 28: Graded as a hosted service. A public Better Stack page at status.navige.ai exists, but its one monitor is `api.trustloop.live/health`, the host of the product's former name, and not api.navige.ai (5 of 20). That monitor shows downtime every day from 26 July to 17 October 2026 (7.221 per cent uptime) under a header saying all services are online, so there is no readable history for the current API (5). No rate-limit numbers in the docs, onboarding guide or terms (0). The docs separate a 503 outage from a 400 bug, state fail-open and fail-closed defaults per action class, return 429 for a blocking budget, pass provider errors through for the client's own retry logic, and make retries count once under `x-navige-transaction`. No Retry-After or backoff guidance (8 of 15). The Terms leave service levels as TBD (0). No beta or preview label on the gateway (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. Schema & documentation, 42 out of 100, up to 9.4 more on the total Why it scored 42: No OpenAPI or other machine-readable contract. `/openapi.json` and `/openapi` return 404 and the docs link none. The MCP endpoint forwards tools from servers the customer registers and defines none of its own (0). No llms.txt on navige.ai or app.navige.ai and no Markdown versions of the docs (0). The docs say what each route does and does not do, for example that `/usage` gates nothing, that Watchtower changes no traffic and that `error` is not `deny` (15 of 20). `/enforce` and `/usage` have field tables with types, a required `tool`, a `classification` enum and 200-character bounds that reject rather than truncate. `arguments` is a free-form object (10 of 15). curl, Python and TypeScript examples on every route, the seven decision shapes, an HTTP status table, five degraded-decision cause codes and a troubleshooting table (14 of 15). No changelog, and `/enforce`, `/usage` and `/mcp` are unversioned paths beside `/v1/transactions` (3 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. ## 3. Security & auth, 53 out of 100, up to 8.2 more on the total Why it scored 53: One `nv_` key per agent, stored as a SHA-256 hash, up to ten per agent, each with last use shown, an optional expiry and revocation that the docs say may take a minute to apply. Keys have no scopes beyond the agent. 25, less 10 because the docs give `?api_key=` in the URL as the MCP route and as the proxy fallback for clients without custom headers (15 of 30). Deny policies, approval holds and autonomy levels L1 to L5 on paid plans, with write, destructive and financial calls failing closed and a kill switch on every tier. On Watchtower, and on `/enforce` generally, the decision is advisory and the caller acts on it (15 of 20). Tool results from registered MCP servers return through Navige to the client. We found no prompt-injection mitigation or guidance (3 of 15). Every call is recorded per agent in an append-only ledger, with signed receipts and optional Polygon anchoring (15). A disclosure policy with a two-business-day acknowledgement and safe harbour, but the security.txt it points to returns 404 on all three hosts, there is no bounty, and DPA 1.2 says Navige holds no certification and has not begun a formal process (5 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. Payments & pricing, 35 out of 100, up to 8.1 more on the total Why it scored 35: No machine payment protocol (0). The pricing page publishes Watchtower at $0 and paid use at the greater of $9 a month or 20 per cent of measured savings, without a login. The Terms still list exact prices, payment terms, taxes and refunds as TBD (15 of 20). Watchtower is free with no card and no time limit, per the pricing page and the Terms (20). A person signs up in a browser and confirms an email, and the Terms forbid creating accounts by automation (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. Agent ergonomics, 61 out of 100, up to 6.3 more on the total Why it scored 61: The proxy returns the provider's response unmodified, decision responses are a few fields, and result capping, tool scoping and context trimming exist to shrink what reaches the agent's context, though all are off by default and paid only (15 of 25). The public API has no list endpoints, since logs and rules moved to the dashboard, and the output ceiling and result caps are the only size controls (8 of 20). Errors are distinct and recoverable, with four 401 reasons, 403 for an inactive or expired agent or the kill switch, 404 for an unknown provider, 429 for a budget and `unavailable: true` on outages (18 of 20). `x-navige-transaction` makes retried model calls count once and `wait: false` avoids a five-minute hold, but `/enforce` and `/usage` take no idempotency key. Navige adds no MCP tools of its own to annotate (10 of 20). One base URL and one header, with only `tool` required on `/enforce`. One official SDK, for Node.js, and the Python SDK named on the security page couldn't be checked (10 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, 50 out of 100, up to 4.4 more on the total Made of editorial 35, provenance 65. Why it scored 50: A closed service with MIT SDKs. The Terms (version 1.0, 13 October 2026) leave prices, liability, warranties, service levels, notice periods and governing law marked TBD (5 of 30). DPA 1.2 (14 October 2026) is specific and corrects version 1.1 in public, stating that no automatic deletion runs, that prompt text through the proxy is not stored and that production data is in the United States. The privacy policy (6 October 2026) contradicts it with per-plan retention of 7 days to 1 year on plans the DPA says no longer exist, and speaks of storing masked prompts (12 of 30). No deprecation policy. The SDK 3.0.0 README lists what was removed, and the DPA promises 30 days' notice of material changes and new deletion schedules, while the Terms leave notice periods TBD (4 of 20). DPA Schedule 2 names Railway (United States), Resend, Stripe and Framer (Netherlands), with transfer safeguards. Google, whose analytics tag the website loads, isn't listed, and locations aren't given for Resend or Stripe (14 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: navige.ai, registered 2026-07-25 (under a year) (0 of 15) - Changelog: not found (0 of 10) - security.txt: not found (0 of 10) ## 7. Maintenance & community, 51 out of 100, up to 4.3 more on the total Why it scored 51: No dated changelog. The latest release found is `navige` 3.0.0 on npm on 7 September 2026, 40 days before this check (20). `navige` 2.0.0 (26 July), 2.0.1 (12 August) and 3.0.0 (7 September) fall within 90 days, as do three releases of `n8n-nodes-navige` (20). A closed service with no public changelog. Support is hello@navige.ai, with a note on each docs page that a person reads reports. The SDK repository's issues weren't readable through the GitHub API from this session (3 of 15). Not in the official MCP registry under either name. The npm SDK covers `intercept()` and approval status only, with 15 methods that now throw, and the Python SDK couldn't be checked (5 of 15). The SDK has no dependencies, and its repository has no CI or tests (3 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. - 2026-10-17. The privacy policy (effective 6 October 2026) says 'Our website and dashboard do not use tracking cookies or third-party analytics.' The navige.ai home, pricing and about pages load Google Analytics (`gtag/js?id=G-G5NVP8M2BB`, configured inline) and Framer's events script, checked first-hand on 17 October 2026. Counted as a misleading claim (-2). The dashboard at app.navige.ai was not checked for analytics (https://app.navige.ai/legal/privacy, https://navige.ai/). ## 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 Python SDK named on the security page. PyPI answered with a bot check. - unchecked: GitHub stars, issues and reply times on SMJAI/trustloop-sdk. The GitHub API was not available from this session, and the repository was read by clone only. - unchecked: the Product Hunt page and its launch date. The lead places the launch on 15 to 17 October 2026, and the home page badge is consistent with that. - unchecked: whether the dashboard at app.navige.ai loads analytics. Only the navige.ai pages were checked. - unchecked: Navige Ltd's registration at Companies House. - Whether the security.txt that the security page points to will be published. It returned 404 on all three hosts. - When the backend rebuild that broke the 2.x SDK took place, and whether customers were told. No changelog or notice was found. - No rate-limit numbers, SLA text, changelog or OpenAPI document were found. ## Weaknesses - The Terms (version 1.0) leave exact prices, payment terms, liability, service levels and governing law marked TBD - No OpenAPI document, llms.txt, changelog or published rate limits were found - The status page monitors only `api.trustloop.live/health`, a retired host shown down since 26 July 2026, while its header says all services are online - The MCP route, and the proxy fallback for clients without custom headers, put the agent key in the URL query string - The privacy policy says the website uses no third-party analytics while navige.ai loads Google Analytics, and its plan retention periods contradict DPA 1.2 ## 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 the Navige key in `x-navige-key` on the proxy. `Authorization` carries the provider key there, and a Navige key in it fails and is forwarded upstream - Act only on `allow` or `approved` from `/enforce`. Treat `error` as neither, and read `unavailable` to tell a malformed request from a Navige outage - Send `x-navige-transaction` with one id per job, reused across that job's retries, so calls group into one transaction and retries are not counted twice - On a 429 from the proxy, check for a blocking budget before retrying. Navige returns 429 when a monthly budget is reached and passes provider errors through without retrying - Expect Claude Code traffic to get measurement, budgets and the kill switch only. Context trimming, routing, the output ceiling and prompt caching are switched off for it ## What the review panel asked for - monitor api.navige.ai - a dated changelog - Fix the Terms' pricing - Publish rate limits - Point status at api.navige.ai - Publish rate limits and an SLA - header-only key on MCP - scoped agent keys ## 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.