{
  "fixes": {
    "slug": "firebase-cloud-messaging",
    "name": "Firebase Cloud Messaging",
    "listing": "https://www.anchorterminal.com/tools/firebase-cloud-messaging",
    "markdown": "# Fix list: Firebase Cloud Messaging\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/firebase-cloud-messaging, the October 2026 research run, assessed 8 October 2026. Grade B, 69.8 out of 100.\n\nThis 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.\n\nFor a coding agent working on Firebase Cloud Messaging: 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.\n\n## 1. Payments \u0026 pricing, 40 out of 100, up to 7.5 more on the total\n\nWhy it scored 40: No x402, MPP or L402 (0). The pricing page lists Cloud Messaging as no-cost on both plans, without a login (20). It is free on the Spark plan, which needs no billing account or card (20). A person signs in with a Google account, accepts the Firebase terms and creates the project and service account in a browser (0).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):\n\nThe published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).\n\n- 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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).\n\nPayment 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.\n\nOpen-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.\n\n## 2. Reliability, 66 out of 100, up to 6.8 more on the total\n\nWhy it scored 66: Hosted reading. Firebase status dashboard with a page and a JSON incident feed per product (20). The feed lists one Cloud Messaging incident between 10 July and 8 October 2026, low availability and higher latency in North America on 1 September from 08:40 to 12:31 UTC, 3 hours 51 minutes, read as one major incident (10). Limits are published with numbers, 600,000 messages a minute per project, 240 a minute and 5,000 an hour to one Android device, 3,000 topic subscription changes a second and 1,000 concurrent fanouts (15). 429 and 503 handling is documented with `retry-after`, a 60 second default, exponential backoff with jitter and a 10 second timeout, but `messages:send` takes no idempotency key and the Admin SDK release notes record timeouts raised to avoid duplicate notifications (11 of 15). The Firebase SLA of 9 April 2020 does not name Cloud Messaging and pays credits as a share of fees, and FCM has no fee, so no SLA (0). The HTTP v1 API is generally available (10).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):\n\nHosted APIs, MCP servers, models and platforms.\n\n- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 10, the surface agents use is generally available, not beta or preview.\n\nLocal packages, SDKs, frameworks and stdio MCP servers.\n\n- 20, installs from an official package with supported runtimes stated.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 15, version 1.0 or later, or declared stable.\n\nProtocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.\n\n## 3. Security \u0026 auth, 68 out of 100, up to 5.6 more on the total\n\nWhy it scored 68: OAuth 2.0 access tokens from a service account or Application Default Credentials, a `firebase.messaging` scope, and IAM permissions per action (30), less 10 because the Discovery document lists `access_token` and `key` as query parameters (20). A custom role can hold `cloudmessaging.messages.create` alone and `validate_only` tests without sending. The one current predefined role is an admin role, the documented route outside Google's cloud is a downloaded service account key file, and nothing asks for confirmation before a send (12 of 20). Responses carry no third-party content (10). Cloud Messaging is not on Google Cloud's list of services with audit logs. Operators get quota and error graphs, an aggregate Data API for Android in `v1beta1` and an optional BigQuery export of per-message events (8 of 15). google.com publishes a security.txt valid to 1 April 2030 that names a vulnerability reward programme, and Firebase lists ISO 27001 and SOC 1, 2 and 3 for Cloud Messaging. firebase.google.com has no security.txt of its own (18 of 20).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-security):\n\n- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.\n\nModels 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.\n\n## 4. Agent ergonomics, 70 out of 100, up to 4.9 more on the total\n\nWhy it scored 70: API reading, graded on the HTTP v1 API. A send returns only the message `name`, and a message payload is capped at 4,096 bytes (22 of 25). Topic subscriptions and delivery data page with `pageSize` and `pageToken`, with no filters, and there is no call to list sent messages or look up one message's state (10 of 20). Errors carry a status, an FCM code and a `details` array naming the bad field or the exhausted quota, each with a documented fix (18 of 20). No idempotency key on send. `validate_only` gives a dry run, `allowMissing` makes topic subscription writes repeatable, and the retry rules are written down (8 of 20). A send needs only a project and one target, and Admin SDKs exist for Node.js, Python, Java, Go and .NET. Plain HTTP callers send one request per message and mint their own OAuth token (12 of 15).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):\n\n- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.\n\nModels 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.\n\n## 5. Schema \u0026 documentation, 86 out of 100, up to 2.3 more on the total\n\nWhy it scored 86: A public Discovery document at `fcm.googleapis.com/$discovery/rest?version=v1`, revision 20261006, with 6 methods and 16 schemas (25). `firebase.google.com/docs/llms.txt` indexes the docs and every page has a `.md.txt` twin (10). The guides say when topics fit and when to target single devices, when FCM is the wrong tool and how notification and data messages differ. The reference itself is terse (16 of 20). 79 typed properties with 4 enums, but `data`, the APNs `payload` and `headers` and the web push `notification` are free-form objects, and `ttl` and `condition` are strings with their own grammar (9 of 15). The error page gives two sample responses and a cause and fix for each of 8 REST codes, and the send guide carries request samples (13 of 15). Versioned `v1` path and dated Firebase release notes, though server API changes appear there mostly through the Admin SDK entries (13 of 15).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):\n\nAPIs and MCP servers.\n\n- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 15, versioning and a public changelog.\n\nModels 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.\n\n## 6. Transparency \u0026 trust, 81 out of 100, up to 1.7 more on the total\n\nMade of editorial 68, provenance 94.\n\nWhy it scored 81: Closed service under the Google APIs Terms of Service and the Firebase Data Processing and Security Terms, with Apache-2.0 Admin SDKs (15 of 30). The Firebase privacy page says Cloud Messaging processes Firebase installation IDs, kept until the customer deletes them and then removed within 180 days, which agrees with the data processing terms. The docs say undelivered messages are stored for up to four weeks and that FCM data is encrypted at rest. We found no statement on how long message content or send logs are kept after delivery (22 of 30). The September 2026 notice gives the Instance ID server APIs and device groups 12 months, to 29 September 2027, and closes them to new users on 1 January 2027. No standing deprecation policy for FCM was found (17 of 20). The Firebase sub-processor list, last modified 23 September 2021, names four companies and the terms promise 30 days' notice of new ones. Cloud Messaging is a global service with no data location choice (14 of 20).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):\n\n- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).\n\nThe other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.\n\nProvenance checks not met in full (half of this category, computed from checked facts):\n\n- Terms of service: read, states 6 of the 7 things a reader expects, and has 1 clause that costs points (7.1 of 10)\n- Privacy policy: read, states 7 of the 8 things a reader expects, and has 1 clause that costs points (7.3 of 10)\n\n## 7. Maintenance \u0026 community, 83 out of 100, up to 1.5 more on the total\n\nWhy it scored 83: Read as a closed service with official SDKs. The Admin SDKs for Node.js and Go shipped on 23 September 2026 and for Java and .NET on 24 September, each moving topic subscriptions to the v1 API, and the Discovery document is revision 20261006 (30). Firebase's release notes carry Cloud Messaging entries on 19, 25 and 27 August and 9, 23 and 24 September 2026 (20). Public release notes, Firebase Support and a bug report form. We could not read the SDK issue trackers, so half marks on answers (10 of 15). Current official Admin SDKs in five languages, `firebase-admin` 14.5.0 on npm and 7.7.0 on PyPI (15). The Node SDK's CI builds and tests on Node 22, 24 and 26, with dependency updates merged on 8 October. We did not see the run results (8 of 10).\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):\n\n- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.\n\nModels are read for deprecation notice periods and model churn rather than release counts.\n\n## What we couldn't check\n\nWhat 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.\n\n- unchecked: GitHub stars and the open issues of the Admin SDK repositories. api.github.com answered with its rate limit, so `githubStars` is empty and issue handling is scored at half\n- unchecked: whether CI passes on the default branch of firebase/firebase-admin-node. We read the workflow file from a clone, not the run results\n- unchecked: the tool definition of `messaging_send_message` in the Firebase MCP server. The listing is graded on the HTTP v1 API\n- The Discovery document lists `access_token` and `key` as query parameters on every method. We took the checklist's 10 points for it, as the Gmail and Google Sheets dossiers did and the Drive and Calendar dossiers did not\n- The Firebase SLA text names no services. We read it as not covering Cloud Messaging because credits are a share of fees and FCM has none\n- No statement was found on how long FCM keeps message content or send logs after delivery\n- The lead was right on the interface. The docs now also describe a v1 topic subscription API and a `fid` target that replaces registration tokens\n\n## Weaknesses\n\n- Push transport only. No in-app feed, email, SMS, user preferences, digests or templates\n- No idempotency key on `messages:send`, so a retried request can reach the device twice\n- One HTTP request per message. Only the Admin SDKs group up to 500 sends in a call\n- FCM is absent from Google Cloud's list of services with audit logs, and per-message delivery data needs a BigQuery export\n- One incident on 1 September 2026 lowered availability in North America for 3 hours 51 minutes, and no SLA names FCM\n\n## What costs an agent a turn today\n\nThe notes we give agents before they call it. Each one is a workaround an agent shouldn't need.\n\n- Mint an access token from a service account with the scope `https://www.googleapis.com/auth/firebase.messaging`, then POST to `https://fcm.googleapis.com/v1/projects/\u003cproject-id\u003e/messages:send`\n- Set `validate_only` to true to test a message or a registration without sending it\n- On 429 wait for the `retry-after` header, or 60 seconds if it is absent. Retry 500 and 503 with exponential backoff and jitter, and never retry 400, 401, 403 or 404\n- Drop a registration when the error is `UNREGISTERED` (404). Keep your own record of sends, because a retry after a timeout can produce a duplicate\n- Target `fid` in new code. The `token` field is marked deprecated in the API, and device groups stop working after 29 September 2027\n\n## When it's done\n\nSend 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.\n",
    "grade": "B",
    "score": 69.8,
    "assessed": "2026-10-08",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 40,
        "maxGain": 7.5,
        "reason": "No x402, MPP or L402 (0). The pricing page lists Cloud Messaging as no-cost on both plans, without a login (20). It is free on the Spark plan, which needs no billing account or card (20). A person signs in with a Google account, accepts the Firebase terms and creates the project and service account in a browser (0).",
        "checklist": [
          "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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-payments"
      },
      {
        "key": "reliability",
        "name": "Reliability",
        "score": 66,
        "maxGain": 6.8,
        "reason": "Hosted reading. Firebase status dashboard with a page and a JSON incident feed per product (20). The feed lists one Cloud Messaging incident between 10 July and 8 October 2026, low availability and higher latency in North America on 1 September from 08:40 to 12:31 UTC, 3 hours 51 minutes, read as one major incident (10). Limits are published with numbers, 600,000 messages a minute per project, 240 a minute and 5,000 an hour to one Android device, 3,000 topic subscription changes a second and 1,000 concurrent fanouts (15). 429 and 503 handling is documented with `retry-after`, a 60 second default, exponential backoff with jitter and a 10 second timeout, but `messages:send` takes no idempotency key and the Admin SDK release notes record timeouts raised to avoid duplicate notifications (11 of 15). The Firebase SLA of 9 April 2020 does not name Cloud Messaging and pays credits as a share of fees, and FCM has no fee, so no SLA (0). The HTTP v1 API is generally available (10).",
        "checklist": [
          "Hosted APIs, MCP servers, models and platforms.",
          "- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 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.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-reliability"
      },
      {
        "key": "security",
        "name": "Security \u0026 auth",
        "score": 68,
        "maxGain": 5.6,
        "reason": "OAuth 2.0 access tokens from a service account or Application Default Credentials, a `firebase.messaging` scope, and IAM permissions per action (30), less 10 because the Discovery document lists `access_token` and `key` as query parameters (20). A custom role can hold `cloudmessaging.messages.create` alone and `validate_only` tests without sending. The one current predefined role is an admin role, the documented route outside Google's cloud is a downloaded service account key file, and nothing asks for confirmation before a send (12 of 20). Responses carry no third-party content (10). Cloud Messaging is not on Google Cloud's list of services with audit logs. Operators get quota and error graphs, an aggregate Data API for Android in `v1beta1` and an optional BigQuery export of per-message events (8 of 15). google.com publishes a security.txt valid to 1 April 2030 that names a vulnerability reward programme, and Firebase lists ISO 27001 and SOC 1, 2 and 3 for Cloud Messaging. firebase.google.com has no security.txt of its own (18 of 20).",
        "checklist": [
          "- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-security"
      },
      {
        "key": "ergonomics",
        "name": "Agent ergonomics",
        "score": 70,
        "maxGain": 4.9,
        "reason": "API reading, graded on the HTTP v1 API. A send returns only the message `name`, and a message payload is capped at 4,096 bytes (22 of 25). Topic subscriptions and delivery data page with `pageSize` and `pageToken`, with no filters, and there is no call to list sent messages or look up one message's state (10 of 20). Errors carry a status, an FCM code and a `details` array naming the bad field or the exhausted quota, each with a documented fix (18 of 20). No idempotency key on send. `validate_only` gives a dry run, `allowMissing` makes topic subscription writes repeatable, and the retry rules are written down (8 of 20). A send needs only a project and one target, and Admin SDKs exist for Node.js, Python, Java, Go and .NET. Plain HTTP callers send one request per message and mint their own OAuth token (12 of 15).",
        "checklist": [
          "- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-ergonomics"
      },
      {
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "score": 86,
        "maxGain": 2.3,
        "reason": "A public Discovery document at `fcm.googleapis.com/$discovery/rest?version=v1`, revision 20261006, with 6 methods and 16 schemas (25). `firebase.google.com/docs/llms.txt` indexes the docs and every page has a `.md.txt` twin (10). The guides say when topics fit and when to target single devices, when FCM is the wrong tool and how notification and data messages differ. The reference itself is terse (16 of 20). 79 typed properties with 4 enums, but `data`, the APNs `payload` and `headers` and the web push `notification` are free-form objects, and `ttl` and `condition` are strings with their own grammar (9 of 15). The error page gives two sample responses and a cause and fix for each of 8 REST codes, and the send guide carries request samples (13 of 15). Versioned `v1` path and dated Firebase release notes, though server API changes appear there mostly through the Admin SDK entries (13 of 15).",
        "checklist": [
          "APIs and MCP servers.",
          "- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-schema"
      },
      {
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 81,
        "maxGain": 1.7,
        "reason": "Closed service under the Google APIs Terms of Service and the Firebase Data Processing and Security Terms, with Apache-2.0 Admin SDKs (15 of 30). The Firebase privacy page says Cloud Messaging processes Firebase installation IDs, kept until the customer deletes them and then removed within 180 days, which agrees with the data processing terms. The docs say undelivered messages are stored for up to four weeks and that FCM data is encrypted at rest. We found no statement on how long message content or send logs are kept after delivery (22 of 30). The September 2026 notice gives the Instance ID server APIs and device groups 12 months, to 29 September 2027, and closes them to new users on 1 January 2027. No standing deprecation policy for FCM was found (17 of 20). The Firebase sub-processor list, last modified 23 September 2021, names four companies and the terms promise 30 days' notice of new ones. Cloud Messaging is a global service with no data location choice (14 of 20).",
        "blend": "editorial 68, provenance 94",
        "checklist": [
          "- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-transparency"
      },
      {
        "key": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 83,
        "maxGain": 1.5,
        "reason": "Read as a closed service with official SDKs. The Admin SDKs for Node.js and Go shipped on 23 September 2026 and for Java and .NET on 24 September, each moving topic subscriptions to the v1 API, and the Discovery document is revision 20261006 (30). Firebase's release notes carry Cloud Messaging entries on 19, 25 and 27 August and 9, 23 and 24 September 2026 (20). Public release notes, Firebase Support and a bug report form. We could not read the SDK issue trackers, so half marks on answers (10 of 15). Current official Admin SDKs in five languages, `firebase-admin` 14.5.0 on npm and 7.7.0 on PyPI (15). The Node SDK's CI builds and tests on Node 22, 24 and 26, with dependency updates merged on 8 October. We did not see the run results (8 of 10).",
        "checklist": [
          "- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.",
          "Models are read for deprecation notice periods and model churn rather than release counts."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-maintenance"
      }
    ],
    "provenance": [
      {
        "label": "Terms of service",
        "value": "read, states 6 of the 7 things a reader expects, and has 1 clause that costs points",
        "points": 7.1,
        "max": 10
      },
      {
        "label": "Privacy policy",
        "value": "read, states 7 of the 8 things a reader expects, and has 1 clause that costs points",
        "points": 7.3,
        "max": 10
      }
    ],
    "unchecked": [
      "unchecked: GitHub stars and the open issues of the Admin SDK repositories. api.github.com answered with its rate limit, so `githubStars` is empty and issue handling is scored at half",
      "unchecked: whether CI passes on the default branch of firebase/firebase-admin-node. We read the workflow file from a clone, not the run results",
      "unchecked: the tool definition of `messaging_send_message` in the Firebase MCP server. The listing is graded on the HTTP v1 API",
      "The Discovery document lists `access_token` and `key` as query parameters on every method. We took the checklist's 10 points for it, as the Gmail and Google Sheets dossiers did and the Drive and Calendar dossiers did not",
      "The Firebase SLA text names no services. We read it as not covering Cloud Messaging because credits are a share of fees and FCM has none",
      "No statement was found on how long FCM keeps message content or send logs after delivery",
      "The lead was right on the interface. The docs now also describe a v1 topic subscription API and a `fid` target that replaces registration tokens"
    ],
    "weaknesses": [
      "Push transport only. No in-app feed, email, SMS, user preferences, digests or templates",
      "No idempotency key on `messages:send`, so a retried request can reach the device twice",
      "One HTTP request per message. Only the Admin SDKs group up to 500 sends in a call",
      "FCM is absent from Google Cloud's list of services with audit logs, and per-message delivery data needs a BigQuery export",
      "One incident on 1 September 2026 lowered availability in North America for 3 hours 51 minutes, and no SLA names FCM"
    ],
    "agentNotes": [
      "Mint an access token from a service account with the scope `https://www.googleapis.com/auth/firebase.messaging`, then POST to `https://fcm.googleapis.com/v1/projects/\u003cproject-id\u003e/messages:send`",
      "Set `validate_only` to true to test a message or a registration without sending it",
      "On 429 wait for the `retry-after` header, or 60 seconds if it is absent. Retry 500 and 503 with exponential backoff and jitter, and never retry 400, 401, 403 or 404",
      "Drop a registration when the error is `UNREGISTERED` (404). Keep your own record of sends, because a retry after a timeout can produce a duplicate",
      "Target `fid` in new code. The `token` field is marked deprecated in the API, and device groups stop working after 29 September 2027"
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-09",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.4",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run"
  }
}