# Fix list: LetsFG From Anchor Terminal's listing at https://www.anchorterminal.com/tools/letsfg, the October 2026 research run, assessed 1 October 2026. Grade B, 64.2 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 LetsFG: 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, 55 out of 100, up to 9 more on the total Why it scored 55: A connector health dashboard at letsfg.co/developers/connectors/health with 1-hour to 7-day windows. It renders client-side, shows connectors rather than the API, and has no incident log (10 of 20). No readable incident history (5). Limits published with numbers, hosted MCP search at 10 per 10 minutes, 30 an hour and 100 a day per user, Developer API `limit` 1 to 200 (15). Over-limit answers 429 with `retry_after_seconds` and escalating pauses, and booking takes an `idempotency_key` with documented retry behaviour (15). No SLA (0). The API, MCP and sandbox aren't labelled 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. Security & auth, 66 out of 100, up to 6 more on the total Why it scored 66: Hosted MCP uses OAuth 2.1 with PKCE (S256), dynamic client registration, five scopes split by search and book for flights and hotels, one-hour access tokens and rotating refresh tokens that revoke the grant on reuse. The Developer API uses an `X-API-Key` header, and we found no key-in-URL option (30). Search and book scopes are separate, the card is held rather than charged until a PNR exists, price changes come back as questions the traveller answers, and changing the card goes through an emailed link (16 of 20). Results come from airline and booking-site pages that LetsFG reads, and we found no injection guidance (5 of 15). `get_agent_profile` shows usage and bookings have status histories; no per-call log for the operator (7 of 15). SECURITY.md with security@letsfg.co and a 48-hour acknowledgement target, CI actions pinned to SHAs, no security.txt (404 per the 30 September check), bug bounty or certification (8 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, 65 out of 100, up to 4.4 more on the total Why it scored 65: The docs describe an MPP lane where a Tempo wallet pays $0.01 once to enrol without a card, offered only when the 402 from /api/agent-access/request carries an `mpp` object. We didn't trigger it, and booking still needs a card (10 of 40). Per-unit prices published without login, $0.01 per flight search past the allowance, $0.005 per hotel search (20). Keyless sandbox that books, a free key with no card, and free MCP search within limits (20). A Developer API key can be registered by API and the sandbox needs no key, but the OAuth consent and card connection need a person in a browser (15 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. Transparency & trust, 67 out of 100, up to 2.9 more on the total Made of editorial 68, provenance 65. Why it scored 67: SDKs, CLI and stdio MCP are MIT. Since June 2026 they're clients of the closed cloud API, which runs under terms naming Mental Balance sp. z o.o. under Polish law (20 of 30). The privacy policy names the controller and gives retention per data type (search sessions 24 months, bookings 3 years after the trip, payment records 5 years, security logs 12 months). No DPA found (25 of 30). Retirements carry dates and the old routes answer 410 naming the replacement, but we found no advance notice period (8 of 20). Processors named (Revolut, Stripe, Google Cloud, Hetzner, Cloudflare, Google Analytics 4) with SCCs for transfers. The AI providers aren't named (15 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: letsfg.co, no registry record we could read (0 of 15) - Changelog: not found (0 of 10) - security.txt: not found (0 of 10) ## 5. Schema & documentation, 83 out of 100, up to 2.8 more on the total Why it scored 83: OpenAPI 3 document (openapi.yaml in the repository, 16 paths, served at /developers/api/openapi.json per the 30 September check) (25). llms.txt at letsfg.co/llms.txt plus Markdown docs, AGENTS.md and SKILL.md in the repository (10). MCP descriptions say when to call a tool and when not to ("do not retry, give the user the booking_url"), but two contradict the docs on cost and limits, see the negative note (14 of 20). Typed inputs with required lists, enums for cabin class and a nested passenger schema; dates are plain strings (12 of 15). An errors page maps every status and code to an action, and the sandbox documents each failure it can simulate (15). CalVer package versions and git tags, and retirements dated in the docs, but no changelog file (7 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. Agent ergonomics, 83 out of 100, up to 2.8 more on the total Why it scored 83: The hosted MCP lists 24 tools, two of them hidden from the model as app-only, and `present_flight_options` cuts results to three picks (18 of 25). Result limits, cabin, time-window and stopover filters, and a `search_id` to page results without re-searching (18 of 20). Error codes an agent can act on, such as 402 `search_allowance_exhausted` versus `payment_method_required`, and `missing_details` with `missing_fields` (20). Idempotency keys on booking with retry guidance; no readOnlyHint or destructiveHint annotations in the stdio server source (12 of 20). Three required search inputs, Python and JavaScript SDKs (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. ## 7. Maintenance & community, 88 out of 100, up to 1.1 more on the total Why it scored 88: Last package release 2026-09-14 (letsfg 2026.5.101 on PyPI, letsfg-mcp 2026.5.77 on npm), commits to 29 September (30). Tags on 30 July, 4 August and 14 September, plus an npm release on 2 September and a registry publication on 6 September (20). Issues #208, #212 and #213 are referenced in fixes on 2 and 11 September, and an outside pull request (#217) was merged on 19 September. Open issue counts unchecked (15 of 25). In the official MCP registry as io.github.LetsFG/letsfg, a GitHub-verified namespace, latest 2026.5.77 (15). CI on push and pull request for both TypeScript packages and the Python SDK, Dependabot for actions. We didn't confirm the branch is green (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. - -4, 2026-09-02 and 2026-09-08. Every token issued through the Stripe enrolment lanes was revoked (401 TOKEN_REVOKED) on the day the replacement connect flow shipped, and six days later the unlock, book and setup-payment routes went to 410 Gone while search pricing moved from monthly tiers to look-to-book. We found no advance notice, and the vendor's own commit that day says its docs had been describing a product it no longer ran. The 410 bodies name the replacement, so we deduct at the lower end. https://github.com/LetsFG/LetsFG/blob/main/docs/api-onboarding.md - -3, current as of 2026-10-01. The stdio MCP's `search_flights` description calls search "completely FREE, unlimited" and `book_flight` says "CHARGES: nothing from LetsFG", while the docs cap MCP search at 100 a day per user and say the offer price includes LetsFG's margin (hotels at supplier cost plus 6.4 per cent, 8.3 per cent on non-EEA cards). https://github.com/LetsFG/LetsFG/blob/main/sdk/mcp/src/index.ts ## 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 the MPP enrolment lane is live; the docs say it appears only when the 402 carries an `mpp` object, and we didn't trigger it - LetsFG's margin on flights, which isn't published - Whether the hosted MCP tools carry readOnlyHint and destructiveHint annotations; the stdio source has none - Whether notice of the September migration went to users by email; none is in the repository or docs - Open GitHub issue counts and reply times, since GitHub's issue pages are closed to our reader ## Weaknesses - Tickets are bought by a LetsFG booking agent after the hold, 4 to 11 minutes per the docs, and bookings can pause with questions - September's Stripe-to-Revolut migration revoked tokens and retired routes with no notice we could find - Tool descriptions call search unlimited and booking free of LetsFG charges, while search is capped and the margin is in the price - No incident history or SLA; the status page shows connector health only - Hotel results take about a minute and refundable hotel cancellations keep 2 per cent ## 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. - Start on `/v1/sandbox/` with no key; the booking states, questions and timings match production - After `book_flight`, poll `get_flight_booking` every 20 to 30 seconds and answer questions before `expires_at_ms`; an unanswered price change fails the booking and releases the hold - Honour `retry_after_seconds` on a 429. Over MCP the limit is 10 searches per 10 minutes and 100 a day, and early retries lengthen the pause - Fetch results with `get_flight_results` and the `search_id` rather than searching again, which spends allowance - Send the same `idempotency_key` when retrying a booking after a timeout - Treat a 410 as a retired route and read the replacement path in its body ## What the review panel asked for - Card rules per lane - Publish the flight margin - Correct the tool descriptions ## 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.