402cron

# Fix list: 402cron

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/402cron, the October 2026 research run, assessed 30 October 2026. Grade B, 63.5 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 402cron: 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, 45 out of 100, up to 11 more on the total

Why it scored 45: Graded as a hosted API. No status page. `GET /health` answers 200 or 503 with the live state (sweep age, the hourly settlement budget, a kill switch, whether the service is selling) and keeps no history, and /status answers 404 (0, and 5 of 30 for the incident record, since no history can be read). Rate limits are published as numbers, 16 payment attempts an hour per caller, 60 settlements an hour service-wide, 10 verifications a day per destination and 40 per client, 50 live tasks and 25 destinations (15). Hourly limits answer 429 with a `retry-after` header and `retryAfterSeconds`, daily ones with `retryTomorrow`, every refusal carries a `retryable` flag where it matters, a replayed settled payment returns the receipt instead of charging again, and the delivery id is stable across attempts for the receiver (15). No SLA; the terms promise only that a request was attempted and cap liability at the credits in question (0). Live and selling, with no beta label, though the earliest public date is 10 September 2026 (10). 0 + 5 + 15 + 15 + 0 + 10 = 45.

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, 41 out of 100, up to 10.3 more on the total

Why it scored 41: One management token per paying wallet, shown once, with every write it can do; no scopes, no expiry, no read-only token. Rotation exists but only as a paid purchase from the same wallet with `?rotate=1`, which kills the old token. The token travels only in the Authorization header (13 of 30). No read-only mode. Pause and resume are separate from PATCH so an edit can't revive a task, a PATCH with an unknown field is refused whole, delete is soft with a 30-day restore, and a destination with a task on it can't be deleted; the three keyless MCP tools are reads (10 of 20). The API returns the agent's own data and delivery outcomes only, and the reply body from the destination isn't stored or returned, so no third-party content reaches the model; the docs tell agents to branch on codes, not prose (10 of 15). A task shows `failsInARow`, `refusedInARow`, `lastRunAt` and `deliveriesLeft`, and deliveries carry `X-402cron-Deliveries-Left`, but no endpoint returns the 90-day attempt log the privacy policy describes (5 of 15). security.txt answers 404, and no disclosure policy, bug bounty or certification was found; an abuse page explains the headers and how to stop deliveries, secrets and task headers are encrypted at rest, and the destination fence refuses private, link-local and internal targets (3 of 20). 13 + 10 + 10 + 5 + 3 = 41.

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. Agent ergonomics, 76 out of 100, up to 3.9 more on the total

Why it scored 76: Graded as an API. Responses are small, task lists take `limit` up to 100, `/docs?section=` fetches one part of the documentation, and the MCP server has 16 tools in one set; no field selection (18 of 25). Cursor pagination with `after` and a `state` filter of 10 values on tasks, an unknown value refused rather than answered with an empty page; the destination list is unpaginated at 25 (17 of 20). Every refusal is `{error, message}` with a stable snake_case code, `limit_reached` names the limit with `max` and `used`, `too_long` names the field and the bytes, `invalid_cron` carries a sub-code, and the one place that answers a sentence is flagged in the docs (19 of 20). A replayed payment nonce returns the receipt, verify and pause are no-ops when repeated, `already_run` stops a second concurrent run, and the delivery id is stable across attempts, but `/run` is charged on every call and `POST /api/tasks` has no idempotency key; MCP annotations weren't read (14 of 20). Three required fields to create a task with method and timeout defaulted, but no official SDK in any language; paying needs a stock x402 client, and the repository's examples are the only client code (8 of 15). 18 + 17 + 19 + 14 + 8 = 76.

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. Maintenance & community, 62 out of 100, up to 3.3 more on the total

Why it scored 62: The last published service change is the starter pack moving to 2,000 deliveries for $1, recorded in the docs repository on 9 October 2026, and the terms and privacy policy took effect on 23 October (30). In the 90 days to 30 October we count registry versions 1.0.0 on 10 September and 1.0.1 on 21 September, the blog post of 19 September, the tariff change of 9 October and the policies of 23 October, dated changes but not a changelog (12 of 20). The docs repository's two open issues, of 23 September and 17 October 2026, have no reply, and support is an email address with no community channel (3 of 25). In the official MCP registry as com.402cron/402cron under the vendor's domain (15). No packages, no CI; the examples pin @x402/core ^2.24.0 (2 of 10). 30 + 12 + 3 + 15 + 2 = 62.

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.

## 5. Transparency & trust, 64 out of 100, up to 3.2 more on the total

Made of editorial 66, provenance 61.

Why it scored 64: The service is closed and its source isn't published; the terms are short and dated, with a liability cap, a 30-day shutdown notice with refund of unused credits and a promise that a tariff change never re-prices held credits; the examples are MIT (17 of 30). The privacy policy of 23 October 2026 gives a retention period for every record it names (attempt log 90 days, usage counters 400 days, payment attempt counters 7 days, failed payments 30 days, settled payments indefinitely, deleted tasks 30 days, unverified destinations 7 days) and agrees with section 7 of the terms; no DPA (25 of 30). A 30-day shutdown notice and a statement that error codes are never renamed, with older response fields kept for compatibility, but no deprecation policy for API changes (10 of 20). Cloudflare Workers, D1 and Durable Objects with a 10 per cent log sample, Coinbase's CDP facilitator and two public Base RPC hosts are named, as is Cloudflare Web Analytics on the public pages; no data location is stated (14 of 20). 17 + 25 + 10 + 14 = 66.

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: 402cron.com, registered 2026-08-22 (under a year) (0 of 15)
- Status page: not found (0 of 10)
- Changelog: not found (0 of 10)
- security.txt: not found (0 of 10)

## 6. Schema & documentation, 81 out of 100, up to 3.1 more on the total

Why it scored 81: A public OpenAPI 3.1 file with 16 paths and 20 operations, one bearer security scheme and schemas for the purchase receipt and every refusal (25). llms.txt and a JSON docs document with `?section=` for agents (10). Descriptions say what to do and what not to, for example that `/run` is charged either way and must not be retried on 200, that HEAD can't settle a payment, that there is no purchase tool over MCP and why; `getTask`, `pause` and `resume` have empty descriptions but their response text carries the rules (17 of 20). Enums on method, task state, overlap and x402 version, minimum and maximum on `limit` and `timeoutMs`, `format: uri` and required fields; `body` is a free string or object by design and `headers` an open map (12 of 15). An example cron and URL, an error catalogue of 57 entries with statuses and sub-codes, and Node and Python receivers plus a client in the repository, but the 200 bodies for task lists and destinations have no schema (12 of 15). `info.version` 1.0.0 and registry versions 1.0.0 and 1.0.1, the docs promise that codes are never renamed and older snake_case fields stay, but there is no changelog page (/changelog answers 404) and no dated list of API changes (5 of 15). 25 + 10 + 17 + 12 + 12 + 5 = 81.

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. Payments & pricing, 86 out of 100, up to 1.8 more on the total

Why it scored 86: x402 v2, with v1 on request, on the service's own pack URLs, USDC on Base through the Coinbase CDP facilitator, documented in the OpenAPI file, /docs, llms.txt and /.well-known/x402, with the client example built on @x402/core. Every sale goes through it. We didn't request the 402 challenge, since the batch's rules forbid calling the API, so 4 are held back for an unprobed flow (36 of 40). Per-attempt prices are public as JSON and on the home page, $0.001 on the trial and $0.0005 on the other packs (20). The trial pack costs $0.02 in USDC rather than being free; no card is involved, and the docs, pricing, health and three MCP tools are free to read (10 of 20). An agent gets the token from the paid 402 exchange with no signup, form or human step (20). 36 + 20 + 10 + 20 = 86.

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.

## 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 live x402 402 challenge and its PAYMENT-REQUIRED header, since the batch's rules forbid calling the API; the payments line rests on the OpenAPI file, /docs, llms.txt and the client example
- unchecked: the MCP tools' input schemas and readOnlyHint or destructiveHint annotations, since no MCP session was opened; the 16 tool names and one-line descriptions come from the paranormalen2/402cron-mcp README
- unchecked: whether the x402 Bazaar lists 402cron. The CDP discovery list had 104,082 entries on 30 October with no search parameter we could use; 5,000 entries at the end of the list held none, and the operator's blog says a resource is dropped after 30 days without a settlement through the CDP facilitator
- unchecked: the GitLab repository named by the registry's 1.0.0 entry (gitlab.com/georgikalchev/402cron), which answered 403 behind a bot check; GitLab's robots.txt disallows its API
- Whether POST /api/tasks upserts on a repeated name or refuses it. The docs call the name the key we upsert on, while name_taken_by_live_task is documented for PATCH and restore only
- What the starter pack held before the 9 October 2026 change; the README commit records only the new size
- Whether a per-second limit applies to the management routes; none is published, and the blog post describes a per-isolate limiter on Workers without numbers for the live service
- It doesn't do LocalGhost's or letme's job

## Weaknesses

- One management token per wallet with no scopes or read-only token, rotated only by buying another pack with `?rotate=1`
- No status page or incident history; /health reports the live state only, and no SLA is published
- No security.txt, disclosure policy, bug bounty or certification was found
- No endpoint returns the delivery log the privacy policy says is kept 90 days; a task shows only counters and the last run time
- The service's first public date is 10 September 2026, the domain was registered on 22 August 2026, and the two GitHub issues on the docs repository have no reply
- Schedules are 5-field cron in UTC at one-minute resolution with no time zones, and a run is skipped while the previous one is still in flight

## 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.

- Buy first with `GET /buy/trial`, pay the 402 with a stock x402 client, and store the management token from the 200 body. It's shown once; losing it costs another purchase with `?rotate=1`.
- Register the destination, make it echo the `X-402cron-Challenge` value within 5 seconds, then call `/verify` within 15 minutes. A fixed-response catch URL can't pass.
- Branch on the `error` code and the HTTP status, never on `message`. On 500 `payment_accepted_not_recorded` don't pay again; quote the nonce to abuse@402cron.com.
- Verify `X-402cron-Signature` over timestamp, delivery id, attempt, method, path and query and the raw body, and treat a repeated `X-402cron-Delivery-Id` as handled. Delivery is at least once.
- Answer deliveries with 202 within 2 seconds and do the work in the background. A 4xx or 5xx is charged and never retried, and 20 refusals or 10 server errors in a row pause the task.
- Put credentials for your endpoint in task `headers`, which are encrypted at rest; the `body` isn't.

## 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.