Hygraph

# Fix list: Hygraph

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/hygraph, the October 2026 research run, assessed 8 October 2026. Grade B, 69.3 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 Hygraph: 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. Payments & pricing, 35 out of 100, up to 8.1 more on the total

Why it scored 35: No x402, MPP or L402 (0). Growth is $199 a month, with overage of $0.20 per 10,000 API operations and per GB of asset traffic. Enterprise is priced through sales (15). The Hobby plan needs no card and includes 500,000 API calls a month (20). A person signs up in a browser by email or social login and creates the token in Studio (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.

## 2. Security & auth, 63 out of 100, up to 6.5 more on the total

Why it scored 63: Permanent Auth Tokens are JWTs sent as a Bearer header, each limited by model, stage, locale, environment and action, with separate Management API permissions, and deleting one invalidates it. No expiry or rotation was found. The global MCP endpoint uses OAuth through auth.hygraph.com with PKCE and a registration endpoint, with access following the user's own permissions (25 of 30). New tokens start with no permissions, MCP token presets include read-only and content-only, MCP rejects delete and unpublish, and public API access is off by default. The API itself has no confirmation step (17). The docs warn against always allowing `execute_graphql`. No guidance on treating stored content as untrusted input was found (5). Audit logs are Enterprise only with 90 days of retention and record the token that made a change (7). The site states SOC 2 Type 2 and ISO 27001 certified hosting. `/.well-known/security.txt` answered 404, no disclosure policy or bounty was found, and the trust report on app.drata.com answered 403 (9).

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, 72 out of 100, up to 4.6 more on the total

Why it scored 72: GraphQL field selection sizes every response, `first` defaults to 10 and is capped at 100, and the MCP tool `list_entities` takes a limit and field selection. The project MCP endpoint lists 17 tools (22 of 25). `first`, `last`, `skip`, `before` and `after`, Relay connections, typed `where` filters and ordering (20). Errors use documented HTTP statuses and a `requestId`, with free-text messages and no machine code (11). No idempotency keys. `upsert` mutations key on a unique field, batch migrations apply all or nothing with a `dry_run` on the MCP tool, and MCP rejects delete and unpublish. We couldn't read the tool annotations (12). Only the Bearer header is required and any GraphQL client works. The one official SDK covers the Management API in JavaScript and TypeScript (7).

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. Schema & documentation, 78 out of 100, up to 3.6 more on the total

Why it scored 78: GraphQL introspection is on by default for each project's Content API, so the contract is machine-readable and specific to the project. The Management API answered 401 to an introspection query without a token, and no static schema file was found (20 of 25). A root `llms.txt` with a Markdown copy of every docs section under `/llms-full/` (10). The Management SDK page states when to use it and when not to, and the MCP tools reference gives each of 19 tools one line with a warning on `execute_graphql` (15). Inputs are typed per model with enums and required fields. Rich Text takes an AST and JSON fields are free-form (13). Request and response examples throughout, and an errors page for nine HTTP statuses. Error bodies carry a message only, with no code list (10). The endpoint path carries `/v2/` and the changelog is public and dated. No API versioning policy was found (10).

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.

## 5. Transparency & trust, 60 out of 100, up to 3.5 more on the total

Made of editorial 45, provenance 75.

Why it scored 60: Closed service under published Terms of Service from Hygraph GmbH, with the Management SDK under MIT (15 of 30). The privacy policy, last updated 19 March 2025, covers the service and lists six categories of service provider without naming them. Retention is stated only as the legal period, the terms disclaim liability for deleted content, and no data processing agreement was found (12). Dated notices for the 2025 move off the legacy asset system and a statement that the old `@graphcms/management` SDK is deprecated. The terms allow the service to change without prior notice (10). The status page and pricing page name 11 hosting regions. Sub-processors are not named (8).

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: hygraph.com, registered 2022-03-04 (4 years) (7 of 15)
- Terms of service: read, states 5 of the 7 things a reader expects, and has 2 clauses that cost points (4.3 of 10)
- Privacy policy: read, states 6 of the 8 things a reader expects (8.5 of 10)
- security.txt: not found (0 of 10)

## 6. Maintenance & community, 74 out of 100, up to 2.3 more on the total

Why it scored 74: The changelog's newest entry is 30 September 2026 (30). Entries on 30 July, 31 August and 30 September, and `@hygraph/management-sdk` 1.6.1 on 19 August (20). Public changelog, a community Slack, in-app chat and email support on Growth. The SDK's GitHub repository shows 9 open issues and no push since 13 September 2024 (9 of 15). The Management SDK is current on npm. The official MCP registry returned no Hygraph entry (11). npm has three SDK releases since June 2026, and the public repository does not carry them, so CI could not be seen (4).

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. Reliability, 89 out of 100, up to 2.2 more on the total

Why it scored 89: Graded on the GraphQL Content and Management APIs, with the hosted MCP server noted. Public status page at status.hygraph.com with the API and the asset CDN as components for 11 regions and an incident feed back to 2021 (20). In the 90 days to 8 October 2026 it lists one incident, 36 minutes on 7 August when new sign-ups were unavailable, marked major, with existing users unaffected (26 of 30). Rate limits published by plan as uncached requests a second (5, 25, up to 500) and concurrent operations (15). The docs give exponential backoff guidance with sample code. `Retry-After` is documented only for asset transformation limits, and there are no idempotency keys (10). The pricing page lists an Enterprise uptime SLA of up to 99.95%. The SLA document was not found (8). The APIs and the MCP server carry no beta label (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.

## 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 trust report linked from hygraph.com/security-and-compliance redirects to app.drata.com, which answered 403, so the SOC 2 report, any penetration test summary and any sub-processor list there were unread
- unchecked: the MCP tool definitions, their JSON Schemas and their readOnlyHint and destructiveHint annotations, because listing tools needs a token. The tool count comes from the docs
- unchecked: whether GitHub CI passes for the Management SDK, because the public repository is behind the npm releases
- No data processing agreement, sub-processor list, SLA document, API versioning policy, token expiry or rotation, or a restore mutation for versions was found in the pages read
- The yearly price for Growth is drawn by script on the pricing page and was not read. Whether failed requests count as API operations was not found
- The lead was right about the vendor and the two GraphQL APIs. It did not mention the hosted MCP server, which the docs list as available on all projects

## Weaknesses

- GraphQL error bodies carry a message and a `requestId` with no machine-readable code, and only asset transformation 429 responses are documented with `Retry-After`
- No idempotency keys in the reviewed documentation. Safe retries rest on `upsert` mutations keyed on a unique field
- Audit logs, backups with recovery and an uptime SLA are Enterprise only, and version history is absent on Hobby and 14 days on Growth
- No data processing agreement, named sub-processor list, security.txt or disclosure policy was found on hygraph.com. The privacy policy names categories of service provider only
- The only official SDK is `@hygraph/management-sdk` for JavaScript and TypeScript, and its public GitHub repository was last pushed on 13 September 2024

## 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 the Permanent Auth Token as `Authorization: Bearer <token>` to `https://<region>.hygraph.com/v2/<projectId>/<environment>`. Read the schema by introspection first, because every type is generated from the project's models
- Mutations write to DRAFT. Call `publish<Model>` with `to: [PUBLISHED]` as a separate mutation, and pass `locales` to write or publish a localisation
- Stay under the plan's limit for uncached requests (5 a second on Hobby, 25 on Growth) and its concurrent mutations (5 and 10). Back off exponentially on 429
- Upload an asset with `createAsset`, then POST the file to the returned pre-signed URL, or pass `uploadUrl` for a remote file. The asset stays `ASSET_CREATE_PENDING` until processed
- Schema changes go to the Management API through `@hygraph/management-sdk` or the MCP tool `submit_batch_migration`. Version restore has no documented mutation, so read `<model>Version` and write the old values back

## 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.