Antseed

# Fix list: Antseed

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/antseed, the October 2026 research run, assessed 8 October 2026. Grade C, 55.1 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 Antseed: 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. Security & auth, 45 out of 100, up to 9.6 more on the total

Why it scored 45: Model reading. The local proxy has no key and binds to 127.0.0.1. One secp256k1 key is identity and wallet, stored in plain text by the CLI unless supplied by environment variable, and encrypted by the OS keychain in the desktop app. Spending is bounded by a deposit contract, a capped ReserveAuth per seller and an optional separate withdrawal wallet, and gateway keys are hashed, revocable and spend-limited (18 of 30). Sellers are independent, and the FAQ says a standard seller can read prompts and sets its own practices (5 of 20). No central log, and an optional TEE-verified route with attestation checked on the buyer's machine (8 of 15). Seller, token and cost headers per response, `antseed buyer metering` and a Prometheus exporter (12 of 15). No security.txt, SECURITY.md, published advisory, bug bounty, certification or contract audit report found. The repository holds one automated review of the buyer proxy dated 8 March 2026 (2 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.

## 2. Reliability, 62 out of 100, up to 7.6 more on the total

Why it scored 62: Read with the local-software lines, since the buyer runs the proxy. Official npm package @antseed/cli with Node.js 20 or later stated, plus desktop installers (20). Public CI with unit tests and an end-to-end payment flow, and the last six runs on main passed on 8 October 2026 (25). 74 open issues and 32 open pull requests against numbering that has reached 1137, with bug reports from late September still open and reply times not sampled (12 of 25). CHANGELOG.md calls out breaking CLI changes, but versions are 0.x patch bumps and everything since August sits under one Unreleased heading (5 of 15). The CLI is 0.1.173 and the release tag v0.2.51, with no statement of stability (0). Whether a request is answered also depends on which sellers are online, which this checklist doesn't score.

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. Schema & documentation, 55 out of 100, up to 7.3 more on the total

Why it scored 55: No OpenAPI document found in the repository or docs. The proxy copies the OpenAI and Anthropic wire formats (0 of 25). llms.txt at antseed.com/llms.txt and a setup skill at /skill.md (10). The API guide states what each route is for, how routing picks a seller and when to pin one (14 of 20). Request bodies follow the upstream formats, and `/v1/models` reports context length, modalities, tool use and supported parameters per seller (8 of 15). Curl examples for each format and an error table with status, code, meaning and fix (13 of 15). A public changelog and tagged releases, with dated entries stopping in August 2026 (10 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.

## 4. Agent ergonomics, 56 out of 100, up to 7.2 more on the total

Why it scored 56: Model reading, as for other routers. Tool use and structured output depend on the seller, and the model list reports both per seller with conservative model-level values (12 of 20 and 8 of 15). Cached-input pricing is advertised per seller where supported (8 of 15). Context length is reported per model as the lowest value across sellers (10 of 15). No batch route found (0 of 10). No SDK of its own for buyers. OpenAI and Anthropic SDKs work against the proxy, and wrappers launch Claude Code, Codex and OpenCode (5 of 10). Documented error codes with fixes, and cost and seller headers on non-streaming responses (13 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.

## 5. Payments & pricing, 70 out of 100, up to 3.8 more on the total

Why it scored 70: x402 v2 exists only on `POST /v1/key/topup` of a gateway the owner runs, and is off until enabled per key. Inference is paid through Antseed's own HTTP 402 flow with EIP-712 signatures and USDC channels on Base, which isn't one of the three protocols the rubric names (10 of 40, a judgement call). Per-token prices for every seller are public JSON without login (20). Free models answer with no card and no account (20). An agent can install the CLI, get a key generated locally and fund it by a USDC transfer, with no sign-up flow (20). The self-hosted rule gives the same three lines for the free software.

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. Transparency & trust, 59 out of 100, up to 3.6 more on the total

Made of editorial 66, provenance 52.

Why it scored 59: GPL-3.0 source for the protocol, CLI, desktop app and contracts (30). The terms, FAQ and privacy page agree that there is no central request log and that sellers may log prompts. There is no privacy policy document, and docs/telemetry.md says the PostHog region and retention period are still to be documented (12 of 30). No deprecation policy found. The changelog gives version cut-offs for protocol changes and a dated cutover of 10 September 2026 for staking contracts (8 of 20). Desktop telemetry is documented event by event with a toggle and an environment switch, and is keyed to the buyer's public wallet address. No telemetry code found in the CLI (16 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: antseed.com, registered 2026-02-23 (under a year) (0 of 15)
- Terms of service: read, states 6 of the 7 things a reader expects (9.1 of 10)
- Status page: not found (0 of 10)
- security.txt: not found (0 of 10)

## 7. Maintenance & community, 79 out of 100, up to 1.8 more on the total

Why it scored 79: v0.2.51 was released on 8 October 2026 and @antseed/cli 0.1.173 was published the same day (30). 28 release tags since 19 August 2026 and 41 CLI versions on npm since 10 July (20). Pull requests merge daily, while 74 issues are open and reply times weren't sampled (13 of 25). The CLI is current on npm, with a node library and ten plugins in the repository, TypeScript only, and no MCP registry entry (8 of 15). CI passes on main, with a pnpm lockfile and a pinned Node version. Electron is pinned to 31.7.6 by an override (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.

- 2026-10-08. The buyer proxy forwards a client's `Authorization`, `x-api-key` and cookie headers to the seller it routes to, confirmed in apps/cli/src/proxy/buyer-proxy.ts at commit 10a1f6b. The repository's own review of 8 March 2026 lists it as VULN-002, and issue #1104 (4 October 2026) is open. Clients started through `antseed claude` or `antseed codex` send a placeholder, so the deduction is 4 of a possible 15. https://github.com/Antseed/antseed/issues/1104

## 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: GitHub's API refused our requests for the rate limit, so reply times on issues and the repository's creation date were not read. Counts and stars come from the web pages
- unchecked: whether a failed or cut-off text request is charged. The docs state it only for video
- unchecked: card top-up fees in the desktop app, and who bears the 4 per cent network fee
- The x402 field is set to unsupported because inference isn't payable by x402. One gateway route, `POST /v1/key/topup`, does speak x402 v2
- Whether sellers listing closed models at a fraction of list price serve the model they name. The snapshot shows the prices, and we ran no requests
- No audit report for the Base contracts was found in the docs or repository. The payments guide says they are verified on BaseScan, which we didn't check
- The Antseed Foundation's jurisdiction and registration aren't stated on the site
- The scout's facts held. One addition, the scout didn't mention that CHANGELOG.md documents x402 on the gateway top-up route

## Weaknesses

- The buyer proxy copies a client's `Authorization` and `x-api-key` headers to the seller. The repository's March 2026 review flagged it and issue #1104 is open
- Sellers are independent and unvetted. The FAQ says a standard seller can read the prompt it serves, and each sets its own data practices
- Pre-1.0 everywhere. The CLI is 0.1.173, the release tag is v0.2.51, and CHANGELOG.md keeps everything since August under one Unreleased heading
- No OpenAPI document, security.txt, SECURITY.md, published advisory, bug bounty or contract audit report was found
- The buyer key is also the wallet, and the CLI stores it in plain text at ~/.antseed/identity.key unless `ANTSEED_IDENTITY_HEX` is set

## 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 a placeholder key such as `antseed-p2p`. The proxy forwards whatever `Authorization` or `x-api-key` header the client sends to the seller
- Call `GET /v1/models` first and request a concrete id. The `antseed` alias works only when the desktop app has a model selected
- Handle 402 `insufficient_deposits` by reading `suggestedAmount`, then fund the address from `antseed buyer deposit` with USDC on Base
- Pass `--verifiers antseed-verifier --require-verifier` to `antseed buyer start` when prompts are private, so only TEE-verified sellers are routed
- Read `x-antseed-peer-id` and `x-antseed-estimated-cost-usd` on each response, and pin a seller with `<peerId>@<model>` when the same one must answer

## 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.