# Fix list: Asmi From Anchor Terminal's listing at https://www.anchorterminal.com/tools/asmi, the October 2026 research run, assessed 14 October 2026. Grade F, 33.3 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 Asmi: 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, 30 out of 100, up to 14 more on the total Why it scored 30: Read with the hosted lines, on the MCP server. No status page found, and status.asmiai.com has no DNS record (0). No readable incident history (5). Plans publish monthly task quotas (20 tasks for $10 and 100 for $49, with search and status unmetered) and no per-call rate limits, the Marmot reading for plan limits with numbers (10 of 15). Tasks run in the background and return an id to poll with execute_status, and the docs say asking for a fresh task link is always safe, with no 429 or backoff guidance (5 of 15). No SLA (0). Nothing labels the service or the server 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. Schema & documentation, 20 out of 100, up to 13 more on the total Why it scored 20: Graded on the MCP server at prod.asmiai.com/mcp/claude/ (also /mcp/openai/ and /mcp/meta/). An unauthenticated initialize returned 401 with a pointer to the OAuth metadata, so the JSON Schemas couldn't be read, and as with Pally the contract line scores as unchecked (0 of 25). No llms.txt or Markdown docs. The site is drawn by script and the docs were read from its JavaScript bundle (0). The MCP docs give each of the 12 tools a one or two sentence purpose in four groups, label each as read-only, acts in the world, changes your account or issues a link, and list what Asmi doesn't do (payments, operating websites). Less than Wajo's tool text, more than Pally's two prose tools (12 of 20). Typed inputs unchecked (0 of 15). Setup steps for Claude, ChatGPT and Muse, a worked task flow and a troubleshooting list naming five failures with a fix for each (a 403 for a token from another client, outreach_not_permitted, a 401 on expiry, expired single-use task links and quota blocks), with no call examples, the Wajo dossier's 8 (8 of 15). No version, release history or changelog for the server (0 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. Agent ergonomics, 43 out of 100, up to 9.3 more on the total Why it scored 43: Graded on the MCP server. Twelve tools per the docs puts it in the 15-point band for 11 to 30 tools. Their definitions are behind OAuth, so size is unchecked, and the Pally dossier took 5 off for that (10 of 25). discover_search returns short results with details fetched separately by id, and tasks_list shows active tasks by default with a status filter, with no paging documented. Wajo's 8 had no filters (10 of 20). The troubleshooting list names five failures, each with its fix, the Wajo reading for named errors without a code list (10 of 20). The docs label each tool's effect, tasks return an id to poll rather than resend, and the dashboard-link tool is marked safe to repeat. Annotations unread and no idempotency key on execute_start. Pally's handle polling got 3 (6 of 20). execute_start takes the goal and Asmi plans the outreach, and there's no SDK (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. ## 4. Security & auth, 49 out of 100, up to 8.9 more on the total Why it scored 49: Graded on the MCP server. OAuth 2.1 with PKCE (S256), dynamic client registration or a client metadata document, two scopes (tasks and outreach), tokens bound to one assistant's endpoint (RFC 8707), 30-day access tokens with rotating refresh tokens stored as hashes, and a revocation endpoint (RFC 7009) in the authorization server metadata read on 14 October 2026. Support also issues 60-day bearer keys for clients without OAuth, revoked by email (27 of 30). Without the outreach scope Asmi only does work between the owner and Asmi, but that scope is ticked by default, and the terms say Asmi won't make a purchase or binding commitment without the approval the product flow requires (12 of 20). Asmi returns web search results and replies from businesses and people, and no prompt-injection guidance was found. The Wajo reading gave 3 where approvals limit what injected text can do, and here the outreach scope is the only such limit (2 of 15). execute_status returns a timeline per task, tasks_list covers every channel and a live dashboard link is available, with no account-wide audit log (8 of 15). No security.txt (the path returns the app shell), and no security contact beyond support@asmiai.com, bug bounty, certification or advisories (0 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. ## 5. Payments & pricing, 30 out of 100, up to 8.8 more on the total Why it scored 30: Read with the hosted rule. No x402, MPP or L402. Upgrades go through a Stripe checkout link that billing_get_payment_link returns, which a person completes (0). Public plans of $10 a month or $99 a year for 20 tasks a month and $49 a month or $499 a year for 100, with no per-task price (10). The MCP docs publish a Free plan (search, task lists and status) and a trial of one task that contacts a third party, and sign-in asks only for a phone number or email and a code (20). Connecting needs a person to finish the sign-in, or a key emailed by support (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. ## 6. Maintenance & community, 3 out of 100, up to 8.5 more on the total Why it scored 3: Read for a closed service with a hosted MCP server. No dated release, changelog entry or app listing found, and the terms and privacy policy carry no dates (0). No dated entries in the last 90 days (0). No changelog, and support at support@asmiai.com with a troubleshooting section, which we didn't test (3 of 15). Not in the official MCP registry (a search for asmi returned only unrelated servers) and no SDK (0). No package or CI to judge (0). 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. ## 7. Transparency & trust, 51 out of 100, up to 4.3 more on the total Made of editorial 43, provenance 58. Why it scored 51: Closed service under terms and a privacy policy that each say Effective October 2026, with no day, naming Humint Labs, Inc. (incorporated in Delaware), 710 Lakeway Drive, Sunnyvale (15). The privacy policy says conversations don't train AI models without explicit consent, keeps call recordings 90 days by default and transcripts and memory for the life of the account, and removes deleted data within 30 days. Stated periods, consistent with the terms, the Natura dossier's 22 (22 of 30). No deprecation policy or dated notices found (2 of 20). The privacy policy's access table names ElevenLabs as the voice infrastructure provider, under a data processing agreement with training on recordings disabled, and Twilio as the telephony provider. Model and hosting providers aren't named and no regions are given. The Tomo dossier gave 4 for providers by category with a few named (4 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: asmiai.com, registered 2025-01-30 (1 year) (3 of 15) - Status page: not found (0 of 10) - Changelog: not found (0 of 10) - security.txt: not found (0 of 10) ## 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 MCP tool JSON Schemas and annotations, which answer only after OAuth. - No day is given for the October 2026 effective date of the terms and privacy policy. - Whether the support-issued static key carries the outreach permission the owner asked for or full access. ## Weaknesses - Tool JSON Schemas and annotations can't be read before OAuth (unauthenticated initialize returns 401). - The outreach scope, which lets Asmi call and text people, is ticked by default on the consent screen. - No status page, changelog, server version or security contact found. - A support-issued 60-day static key lets whoever holds it act as the owner. - Model and hosting providers aren't named in the privacy policy. ## 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. - Use the endpoint for your own client (/mcp/claude/, /mcp/openai/ or /mcp/meta/). A token from another client gets 403. - Call discover_search before execute_start, then poll execute_status with the task id instead of starting the task again. - Answer Asmi's questions with execute_advance. Calls already placed can't be recalled by execute_cancel. - Expect outreach_not_permitted if the owner unticked contact people. Ask them to reconnect rather than retrying. - Check billing_status before starting a task, since each outreach task uses one of the month's quota. ## 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.