# Fix list: Pomerado From Anchor Terminal's listing at https://www.anchorterminal.com/tools/pomerado, the October 2026 research run, assessed 17 October 2026. Grade B, 62.6 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 Pomerado: 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, 40 out of 100, up to 12 more on the total Why it scored 40: Graded as a hosted service, Pomerado Cloud, since the MCP and REST docs and the homepage's setup prompt point agents to it. No status page found. status.pomerado.ai has no DNS record and none is linked from the site or docs (0). No readable incident history (5). Daily allowances are published with numbers (100 or 1,000 tool runs a day, 10 or 100 integrations a month), keyless runs have a daily limit, and 20 connect keys an hour, but no per-second or per-minute request rate was found (10 of 15). Over a limit the API answers 429 `quota_exceeded` with `reset_at`, or `try_limit_reached` with Retry-After, every error carries `retryable`, retryable MCP errors add `retry_after_seconds`, and runs, builds and invitations take idempotency keys with `write_status` and `outcome_unknown` for writes whose result is unclear (15). Business lists SLA guarantees, with no published terms (0). No beta or preview label on the hosted MCP or API, and the registry entry is version 1.0.0 (10). The service is 12 days old by its first npm release. 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. Payments & pricing, 42 out of 100, up to 7.2 more on the total Why it scored 42: No x402, MPP or L402 on Pomerado's endpoints (0). Plan prices are public, Free $0, Premium $50 a month and Business custom, with call and build allowances, but there is no per-call price (10). The Free plan and keyless use need no card (20). An agent can find and run 46 public read-only tools with no account, but writes, logins and builds need an account that a person approves, through OAuth or a device code, so partial credit (12 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. ## 3. Security & auth, 69 out of 100, up to 5.4 more on the total Why it scored 69: OAuth for MCP clients, and `pom_` API keys limited to chosen permissions out of 11, expiring after 90 days by default, rotated and revoked through the API, with integration keys confined to one site's MCP. No secret travels in a URL; the keyless URL refuses a key (30). The read-only run refuses a tool that writes before a job starts, builds declare read or write, write tools carry destructiveHint, website logins go through a protected page, and a key can't reveal a saved password or authenticator code. No approval step of Pomerado's own for writes was found beyond the client's own prompt (16). Tools return website content. The job-object page says to treat a job's `message` as data, never as instructions, and outputs are checked against each tool's schema, but no wider guidance on untrusted page text was found (6). Run logs list each browser action and the site's own requests, kept 90 days, builds have an activity log, and connected apps show permissions used and last use (13). No security.txt, SOC 2, ISO 27001 or bug bounty found. The README sends reports to GitHub private vulnerability reporting, and the terms commit to notify customers of incidents without undue delay (4). Pomerado stores website passwords, authenticator secrets and recovery codes, which raises the stakes of that gap. 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. Transparency & trust, 54 out of 100, up to 4 more on the total Made of editorial 43, provenance 65. Why it scored 54: The hosted service is closed with clear terms (Common Paper Cloud Service Agreement 2.1, last updated 4 October 2026), and the complete minting core is MIT on GitHub, which the README says the Cloud runs (24). The privacy policy is an iubenda embed whose path iubenda's robots.txt disallows, so it was not read. The homepage FAQ says results are erased once read or within 60 minutes, credentials are kept only if saved, and hashes of sign-in identifiers are retained, and the API docs keep run logs for 90 days. No public DPA found (10 of 30). No deprecation policy found. The tool docs promise that a repair never removes or loosens an output field unless the site stopped showing it and the person confirms, and the terms give 30 days' notice of price changes (7). No subprocessor list or data location found. The open-source README says builds use OpenAI models (2). 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: pomerado.ai, registered 2026-08-28 (under a year) (0 of 15) - Status page: not found (0 of 10) - security.txt: not found (0 of 10) ## 5. Agent ergonomics, 83 out of 100, up to 2.8 more on the total Why it scored 83: The keyless endpoint lists eight tools with about 24,000 characters of definitions, and the docs give 14 for a signed-in connection, with the rest of the API reached through `search_pomerado_api`, `describe_pomerado_api` and `call_pomerado_api` instead of one tool per operation. An integration MCP narrows an agent to one site's tools (19 of 25). Lists take `cursor` and `limit` 1 to 100, tool search filters by query, site and integration, `get_job` takes `include_result` and `wait_seconds`, and paged results carry a cursor (17). One error envelope with `code`, `message`, `retryable`, `docs_url` and `details`, and codes such as `key_permission_missing` name the missing permission and the page to fix it (20). Idempotency keys on runs and builds, with replay marked `Idempotent-Replayed`, and every keyless tool carries readOnlyHint or destructiveHint, idempotentHint and openWorldHint as fits (20). Few required fields (none for search, two for a run), but no official client SDK. The npm package is the local minter, not a Cloud client (7 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. Schema & documentation, 86 out of 100, up to 2.3 more on the total Why it scored 86: OpenAPI 3.1 at /v1/openapi.json with 84 operations on 66 paths, and every MCP tool on the keyless endpoint has a typed JSON Schema input with `additionalProperties: false` (25). llms.txt with every page, and a Markdown twin of each docs page (10). MCP descriptions say when to use each tool and when not to, such as Pomerado before a browser, the read-only run for read tools, `possible_matches` before a build, and secrets on `answer_url`, not in chat. The server's instructions say to use a browser only if Pomerado says it can't build a tool (18). 159 enums, lengths and ranges such as `limit` 1 to 100 and keys of 1 to 200 characters, and required fields marked. The run and build tools take `input` and `example_input` as JSON-encoded strings over MCP (11). 58 of 84 operations carry examples, all 84 document error responses, and the errors page gives 132 codes, each with its status, whether to retry and what to do (14). Paths are versioned `/v1`. The API reference has a dated Changes list with one entry, of 15 October 2026, and the open-source core keeps CHANGELOG.md with breaking-change sections (8 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. ## 7. Maintenance & community, 81 out of 100, up to 1.7 more on the total Why it scored 81: npm 0.4.0-canary.194 on 17 October 2026, the API change of 15 October and the registry entry of 15 October (30). Releases 0.1.0 to 0.2.0 between 5 and 6 October and more than 150 canaries since (20). About 650 merged pull requests in 12 days, nearly all by the two founders. GitHub issues and replies couldn't be read, because the GitHub API refused this session, so this rests on the commit history and the founders' email address (8 of 25). In the official MCP registry as `ai.pomerado/pomerado`, a domain-verified namespace, active since 15 October 2026 (15). CI runs a gitleaks secret scan, pins actions by SHA and publishes with npm provenance. Run results couldn't be read (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. - Misleading claim, -2. On 17 October 2026 the homepage (https://pomerado.ai) says "Integrations for all sites" and "If a person can use the site, your agent can use it too", and lists bank portals and government filings among supported sites. The docs say builds for bank and government websites are refused with `site_not_supported`, `details.category` `bank` or `government` (https://docs.pomerado.ai/errors#site_not_supported, https://docs.pomerado.ai/guides/build/overview). ## 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. - Judgement call on category. Listed in browser, not web-scrapers, because its tools act on websites (carts, bookings, forms) as well as read them, and its nearest listings, Skyvern and TinyFish, are there. Its agent never drives the browser itself, so `browser.control` stands for the browser work Pomerado does for it. - Graded on Pomerado Cloud, the hosted MCP and REST API, which the homepage and docs send agents to. The MIT core on npm is a local stdio MCP server that needs the owner's OpenAI key and is named in `details`, not graded separately. - The lead dated the launch to 15 to 17 October. The first first-hand dates are the npm package and the repository's initial commit on 5 October 2026, so `firstReleased` is 2026-10-05. The hosted MCP entered the official MCP registry on 15 October, and the docs sitemap was last modified on 9 October. The Product Hunt launch was not read. - Not one of our own products. `find_website_tool` searches only Pomerado's own catalogue of tools it built, and it doesn't pick between other vendors' tools by grades, which is letme's job. It isn't a personal AI platform like LocalGhost. - unchecked: the privacy policy, because iubenda's robots.txt disallows /privacy-policy/ where it is hosted. - unchecked: GitHub issues, pull request discussions, CI run results, private vulnerability reporting settings and the star count, because the GitHub API refused this session. The repository was read by git clone. - unchecked: the signed-in Pomerado MCP's 14 tools and their annotations, which we took from the docs because we made no account. The keyless endpoint's eight tools were listed first-hand. - We didn't run a tool, so the vendor's claims of speed, deterministic results, PII redaction and automatic repair are not measured here. ## Weaknesses - First released on 5 October 2026, with a team of two per its YC profile - No status page or per-second rate limits, and the SLA guarantees listed for Business have no published terms - No security.txt, SOC 2 report or bug bounty found, while it stores website passwords, authenticator secrets and recovery codes - The homepage lists bank portals and government filings, but builds for bank and government sites are refused with `site_not_supported` - Run tools take `input` as a JSON-encoded string over MCP, and there is no official client SDK ## 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. - Try public read-only tools at https://mcp.pomerado.ai/mcp/keyless, then use `sign_in` and reconnect at /mcp with the key for writes, logins and builds - Call `find_website_tool` with a query and `site_url`, and check `possible_matches` before starting a build - Send an `idempotency_key` with every run that writes, and on `outcome_unknown` read the website back before calling again - Give the person the job's `answer_url` for website logins and codes. Don't take passwords in chat - Save a run's result when it arrives. It is returned once and erased after 60 minutes if unread ## What the review panel asked for - longer result retention - Publish per-run overage prices - Publish the keyless limit - homepage that matches the refusal list - publish a SOC 2 report - server-side write confirmation ## 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.