{
  "fixes": {
    "slug": "collibra",
    "name": "Collibra",
    "listing": "https://www.anchorterminal.com/tools/collibra",
    "markdown": "# Fix list: Collibra\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/collibra, the October 2026 research run, assessed 9 October 2026. Grade B, 64.6 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 Collibra: 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, 0 out of 100, up to 12.5 more on the total\n\nWhy it scored 0: Hosted service, so the hosted rubric applies. No x402, MPP or L402 (0). No published prices. www.collibra.com/pricing returns 404 and the site links a demo request and a product tour (0). No free tier or self-serve trial was found. The Collibra Units included in a contract are an allowance inside a paid subscription (0). An agent needs a customer's administrator to register an OAuth application after a sales contract (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 service, read on the hosted lines and graded on the REST and GraphQL APIs of a Collibra cloud instance, with the MCP server as a second surface. status.collibra.com lists six components for each of four regions with an incident history (20). In the 90 days to 9 October 2026 the page shows a lineage and integrations disruption from 24 July to 14 August in all regions, where two ingestion methods created duplicate mapping IDs, and the close on 5 August of a UAE region disruption that began on 1 March. The page reported all systems normal for the 55 days before our check (15 of 30). No general request rate limit was found. The OAuth settings page gives a pool of 100 token requests for each client refilled at 30 a minute, Knowledge Graph queries stop at 100,000 nodes, and 8 imports run concurrently (7 of 15). A 429 is documented for the token endpoint with its causes, and the Import API guide says to retry a job that fails on a concurrency error. No `Retry-After` header or idempotency key was found (6 of 15). The Service Level Agreement of 1 October 2023 sets a 99.5 per cent monthly uptime target with service credits (10). The REST APIs are generally available. The hosted MCP documentation states no status, and the local server is at v0.0.62 (8 of 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 client credentials for integration applications, which act as an application-specific system user, and an authorisation code flow with PKCE for MCP clients. Tokens are valid for 5 minutes. The OAuth settings page says Collibra assigns all scopes to the client, so permissions follow Collibra roles. HTTP Basic with a username and password remains the first method in the tutorials and the only one in the local MCP server (22 of 30). The MCP server respects the signed-in account's permissions, names the permission each tool needs, keeps write-heavy data quality tools off by default and asks for `confirm=true` before a delete. `create_asset` and `edit_asset` write without a confirmation step (15 of 20). Tools return descriptions, comments and definitions written by catalogue users, and we found no prompt-injection guidance (3 of 15). An Activities API and asset history record changes, OAuth actions are attributed to the application's system user, and the Security Policy keeps logs for up to a year (11 of 15). SOC 1, SOC 2, ISO 27001, 27017, 27018 and 42001, FedRAMP Moderate and TISAX per the trust centre, annual third-party penetration tests, and a bug bounty paying 10 to 500 US dollars. `security.txt` expired on 31 December 2025 and no public advisory page was found (17 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, 75 out of 100, up to 4.1 more on the total\n\nWhy it scored 75: The local MCP server registers 30 tools by default and keeps 18 data quality tools behind a flag, with allow and deny lists for tool names. GraphQL queries select fields, and REST lists take `limit` (21 of 25). REST paging is by `offset` and `limit` or by `cursor` with `nextCursor`, with filters on most list operations and a `countLimit` switch to skip the total. GraphQL takes `where`, `order`, `limit` and `offset` (19 of 20). Errors are standard HTTP codes with short descriptions, and the MCP source returns messages that list the valid options when a name does not resolve. No catalogue of API error codes was found (13 of 20). Every MCP tool declares `readOnlyHint`, `destructiveHint` and `idempotentHint`, and the destructive data quality tools return a preview until called with `confirm=true`. The Import API matches resources by identifier, so a repeated import updates in place. No idempotency keys on the REST APIs (15 of 20). Defaults are sensible (20 search results, 10 GraphQL items). We found no official SDK. The portal has a tutorial on generating a client from the OpenAPI definition, and a CLI binary with two command groups (7 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, 83 out of 100, up to 2.8 more on the total\n\nWhy it scored 83: Each REST API has an OpenAPI 3.0 definition, printed in the Markdown copy of every reference page and served by each instance at `/docs/index.html`, and the GraphQL schema is open to introspection. MCP tools take typed inputs and declare an output schema. We found no single public file to download (22 of 25). developer.collibra.com publishes `llms.txt`, a Markdown copy of each page and an MCP server for the documentation (10). MCP tool descriptions state purpose, required permissions and the tool to call next, and the server ships an instructions file. REST operations carry a one-line summary (15 of 20). Request schemas use enums, defaults, minimum and maximum values and required fields. The Output Module takes a query document of its own (13 of 15). Tutorials cover authentication, pagination and import with examples. Error responses are generic descriptions for 400, 401, 403, 404 and 500, plus a table of four JWT error codes (10 of 15). Paths carry a version (`/rest/2.0`, `/v1`), release notes appear monthly and the Output Module guide keeps a dated list of changes. The API compatibility page showed our reader no content (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. Maintenance \u0026 community, 76 out of 100, up to 2.1 more on the total\n\nWhy it scored 76: The release calendar shows 2026.09 reaching commercial production on 4 October 2026, and the local MCP server tagged v0.0.62 on 5 October (30). Releases 2026.07, 2026.08 and 2026.09 reached production on 26 July, 30 August and 4 October, and the MCP repository tagged 15 versions between 11 August and 5 October (20). Closed service, scored on the 15-point scale. Monthly release notes, an announcements page, a community site and a support policy with a 2-hour first response for urgent incidents. The MCP repository's 19 open items on GitHub are all pull requests, most from Collibra staff (11 of 15). No official SDKs and no entry in the official MCP registry. The CLI was last tagged v2.1.1 on 10 August 2026 (7 of 15). The MCP repository runs tests with the race detector and a linter on every push, uses Renovate and builds on Go 1.25 with the official MCP Go SDK 1.7.0 (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## 7. Transparency \u0026 trust, 82 out of 100, up to 1.6 more on the total\n\nMade of editorial 75, provenance 89.\n\nWhy it scored 82: Closed service under the published Master Collibra Agreement and Developer Terms. The local MCP server is Apache-2.0 and the CLI ships under a Binary Code Licence Agreement (15 of 30). The agreement allows export for 30 days after the term and then deletion on Collibra's standard schedule, and the Data Processing Addendum of 1 October 2026 sets deletion within 120 days of termination and breach notice without undue delay. The agreement bars sending customer data to a third party to train generative AI without written consent. The Privacy Policy says it does not cover data processed for customers. No retention period for a live subscription was found (24 of 30). An announcements page gives end-of-life dates, an 18-month deprecation period for retired REST properties and removal of the Metric feature in 2027.01. The API tutorial says breaks happen only in a major release after a deprecation message, and a supported versions policy is published (17 of 20). The subprocessor list of 8 September 2026 names each processor with its purpose and hosting location, and the addendum gives 30 days' notice of additions (19 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: published, but our reader couldn't read it (7 of 10)\n- Privacy policy: published, but our reader couldn't read it (7 of 10)\n- security.txt: published but past its Expires date (5 of 10)\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- The lead said the only MCP server serves the documentation. Collibra also documents a hosted catalogue MCP server at `/rest/mcp` and publishes a local one, and both are graded here as a second surface.\n- unchecked: the API compatibility page at productresources.collibra.com/docs/collibra/latest/Content/ReleaseNotes/ref_compatibility-api.htm showed only navigation to our reader, so the list of API changes by release is unread.\n- No general REST or GraphQL request rate limit and no `Retry-After` guidance were found in the developer portal or the product documentation pages we read. A limit may be documented on a page we did not open.\n- Whether the hosted MCP server is generally available or in preview is not stated on the pages read. The overview page is dated 26 June 2026.\n- The hosted MCP server's tool definitions were read from the documentation table and the open-source repository, not from a live `tools/list`, so the count on a given instance may differ.\n- No self-serve trial was found. The subprocessor list names a vendor for trial and beta testing and an Evaluation Agreement is published, so evaluations may exist through sales.\n- The Developer Terms and the Master Collibra Agreement forbid publishing benchmarks or performance information, and the Developer Terms forbid testing the toolkit's capabilities. This matters before our probes run.\n- The legal documents are PDFs behind cover pages, which the site's terms and privacy reader may not fetch.\n- One request went to rdap.verisign.com before its robots.txt was fetched. The robots.txt request answered 400, so nothing was disallowed.\n- The year Collibra first released the platform was not established from the pages read, so `firstReleased` is null.\n\n## Weaknesses\n\n- No published price, trial or self-serve sign-up. The pricing address returns 404 and the site links a demo request and a product tour\n- No general request rate limit or `Retry-After` guidance found. Only the OAuth token endpoint has numbers, a pool of 100 tokens refilled at 30 a minute\n- The Developer Terms of 22 August 2023 forbid publishing benchmarks and calls not driven by bona fide end user requests, except reasonable testing\n- No prompt-injection guidance found, although tools return descriptions and comments written by catalogue users\n- The local MCP server signs in with a Collibra username and password only, and `security.txt` expired on 31 December 2025\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- Ask a Collibra administrator to register an OAuth application. Integration applications use client credentials, and MCP clients need a user application with a redirect URI\n- Reuse each access token for its 5 minutes. The token endpoint answers 429 once a client drains its pool of 100 tokens, refilled at 30 a minute\n- Page large reads with `cursor=` and `sortField=ID`. Cursor and offset parameters cannot be combined, and Search API `limit` plus `offset` cannot exceed 10,000\n- Keep Import API files to 10,000 assets or fewer and run at most 8 imports at once. Retry a job that fails on a concurrency error\n- Expect `discover_data_assets` and `discover_business_glossary` to consume Collibra Units and to need the `dgc.ai-copilot` permission. Use `search_asset_keyword` for free keyword search\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": 64.6,
    "assessed": "2026-10-09",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 0,
        "maxGain": 12.5,
        "reason": "Hosted service, so the hosted rubric applies. No x402, MPP or L402 (0). No published prices. www.collibra.com/pricing returns 404 and the site links a demo request and a product tour (0). No free tier or self-serve trial was found. The Collibra Units included in a contract are an allowance inside a paid subscription (0). An agent needs a customer's administrator to register an OAuth application after a sales contract (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 service, read on the hosted lines and graded on the REST and GraphQL APIs of a Collibra cloud instance, with the MCP server as a second surface. status.collibra.com lists six components for each of four regions with an incident history (20). In the 90 days to 9 October 2026 the page shows a lineage and integrations disruption from 24 July to 14 August in all regions, where two ingestion methods created duplicate mapping IDs, and the close on 5 August of a UAE region disruption that began on 1 March. The page reported all systems normal for the 55 days before our check (15 of 30). No general request rate limit was found. The OAuth settings page gives a pool of 100 token requests for each client refilled at 30 a minute, Knowledge Graph queries stop at 100,000 nodes, and 8 imports run concurrently (7 of 15). A 429 is documented for the token endpoint with its causes, and the Import API guide says to retry a job that fails on a concurrency error. No `Retry-After` header or idempotency key was found (6 of 15). The Service Level Agreement of 1 October 2023 sets a 99.5 per cent monthly uptime target with service credits (10). The REST APIs are generally available. The hosted MCP documentation states no status, and the local server is at v0.0.62 (8 of 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 client credentials for integration applications, which act as an application-specific system user, and an authorisation code flow with PKCE for MCP clients. Tokens are valid for 5 minutes. The OAuth settings page says Collibra assigns all scopes to the client, so permissions follow Collibra roles. HTTP Basic with a username and password remains the first method in the tutorials and the only one in the local MCP server (22 of 30). The MCP server respects the signed-in account's permissions, names the permission each tool needs, keeps write-heavy data quality tools off by default and asks for `confirm=true` before a delete. `create_asset` and `edit_asset` write without a confirmation step (15 of 20). Tools return descriptions, comments and definitions written by catalogue users, and we found no prompt-injection guidance (3 of 15). An Activities API and asset history record changes, OAuth actions are attributed to the application's system user, and the Security Policy keeps logs for up to a year (11 of 15). SOC 1, SOC 2, ISO 27001, 27017, 27018 and 42001, FedRAMP Moderate and TISAX per the trust centre, annual third-party penetration tests, and a bug bounty paying 10 to 500 US dollars. `security.txt` expired on 31 December 2025 and no public advisory page was found (17 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": 75,
        "maxGain": 4.1,
        "reason": "The local MCP server registers 30 tools by default and keeps 18 data quality tools behind a flag, with allow and deny lists for tool names. GraphQL queries select fields, and REST lists take `limit` (21 of 25). REST paging is by `offset` and `limit` or by `cursor` with `nextCursor`, with filters on most list operations and a `countLimit` switch to skip the total. GraphQL takes `where`, `order`, `limit` and `offset` (19 of 20). Errors are standard HTTP codes with short descriptions, and the MCP source returns messages that list the valid options when a name does not resolve. No catalogue of API error codes was found (13 of 20). Every MCP tool declares `readOnlyHint`, `destructiveHint` and `idempotentHint`, and the destructive data quality tools return a preview until called with `confirm=true`. The Import API matches resources by identifier, so a repeated import updates in place. No idempotency keys on the REST APIs (15 of 20). Defaults are sensible (20 search results, 10 GraphQL items). We found no official SDK. The portal has a tutorial on generating a client from the OpenAPI definition, and a CLI binary with two command groups (7 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": 83,
        "maxGain": 2.8,
        "reason": "Each REST API has an OpenAPI 3.0 definition, printed in the Markdown copy of every reference page and served by each instance at `/docs/index.html`, and the GraphQL schema is open to introspection. MCP tools take typed inputs and declare an output schema. We found no single public file to download (22 of 25). developer.collibra.com publishes `llms.txt`, a Markdown copy of each page and an MCP server for the documentation (10). MCP tool descriptions state purpose, required permissions and the tool to call next, and the server ships an instructions file. REST operations carry a one-line summary (15 of 20). Request schemas use enums, defaults, minimum and maximum values and required fields. The Output Module takes a query document of its own (13 of 15). Tutorials cover authentication, pagination and import with examples. Error responses are generic descriptions for 400, 401, 403, 404 and 500, plus a table of four JWT error codes (10 of 15). Paths carry a version (`/rest/2.0`, `/v1`), release notes appear monthly and the Output Module guide keeps a dated list of changes. The API compatibility page showed our reader no content (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": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 76,
        "maxGain": 2.1,
        "reason": "The release calendar shows 2026.09 reaching commercial production on 4 October 2026, and the local MCP server tagged v0.0.62 on 5 October (30). Releases 2026.07, 2026.08 and 2026.09 reached production on 26 July, 30 August and 4 October, and the MCP repository tagged 15 versions between 11 August and 5 October (20). Closed service, scored on the 15-point scale. Monthly release notes, an announcements page, a community site and a support policy with a 2-hour first response for urgent incidents. The MCP repository's 19 open items on GitHub are all pull requests, most from Collibra staff (11 of 15). No official SDKs and no entry in the official MCP registry. The CLI was last tagged v2.1.1 on 10 August 2026 (7 of 15). The MCP repository runs tests with the race detector and a linter on every push, uses Renovate and builds on Go 1.25 with the official MCP Go SDK 1.7.0 (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"
      },
      {
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 82,
        "maxGain": 1.6,
        "reason": "Closed service under the published Master Collibra Agreement and Developer Terms. The local MCP server is Apache-2.0 and the CLI ships under a Binary Code Licence Agreement (15 of 30). The agreement allows export for 30 days after the term and then deletion on Collibra's standard schedule, and the Data Processing Addendum of 1 October 2026 sets deletion within 120 days of termination and breach notice without undue delay. The agreement bars sending customer data to a third party to train generative AI without written consent. The Privacy Policy says it does not cover data processed for customers. No retention period for a live subscription was found (24 of 30). An announcements page gives end-of-life dates, an 18-month deprecation period for retired REST properties and removal of the Metric feature in 2027.01. The API tutorial says breaks happen only in a major release after a deprecation message, and a supported versions policy is published (17 of 20). The subprocessor list of 8 September 2026 names each processor with its purpose and hosting location, and the addendum gives 30 days' notice of additions (19 of 20).",
        "blend": "editorial 75, provenance 89",
        "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"
      }
    ],
    "provenance": [
      {
        "label": "Terms of service",
        "value": "published, but our reader couldn't read it",
        "points": 7,
        "max": 10
      },
      {
        "label": "Privacy policy",
        "value": "published, but our reader couldn't read it",
        "points": 7,
        "max": 10
      },
      {
        "label": "security.txt",
        "value": "published but past its Expires date",
        "points": 5,
        "max": 10
      }
    ],
    "unchecked": [
      "The lead said the only MCP server serves the documentation. Collibra also documents a hosted catalogue MCP server at `/rest/mcp` and publishes a local one, and both are graded here as a second surface.",
      "unchecked: the API compatibility page at productresources.collibra.com/docs/collibra/latest/Content/ReleaseNotes/ref_compatibility-api.htm showed only navigation to our reader, so the list of API changes by release is unread.",
      "No general REST or GraphQL request rate limit and no `Retry-After` guidance were found in the developer portal or the product documentation pages we read. A limit may be documented on a page we did not open.",
      "Whether the hosted MCP server is generally available or in preview is not stated on the pages read. The overview page is dated 26 June 2026.",
      "The hosted MCP server's tool definitions were read from the documentation table and the open-source repository, not from a live `tools/list`, so the count on a given instance may differ.",
      "No self-serve trial was found. The subprocessor list names a vendor for trial and beta testing and an Evaluation Agreement is published, so evaluations may exist through sales.",
      "The Developer Terms and the Master Collibra Agreement forbid publishing benchmarks or performance information, and the Developer Terms forbid testing the toolkit's capabilities. This matters before our probes run.",
      "The legal documents are PDFs behind cover pages, which the site's terms and privacy reader may not fetch.",
      "One request went to rdap.verisign.com before its robots.txt was fetched. The robots.txt request answered 400, so nothing was disallowed.",
      "The year Collibra first released the platform was not established from the pages read, so `firstReleased` is null."
    ],
    "weaknesses": [
      "No published price, trial or self-serve sign-up. The pricing address returns 404 and the site links a demo request and a product tour",
      "No general request rate limit or `Retry-After` guidance found. Only the OAuth token endpoint has numbers, a pool of 100 tokens refilled at 30 a minute",
      "The Developer Terms of 22 August 2023 forbid publishing benchmarks and calls not driven by bona fide end user requests, except reasonable testing",
      "No prompt-injection guidance found, although tools return descriptions and comments written by catalogue users",
      "The local MCP server signs in with a Collibra username and password only, and `security.txt` expired on 31 December 2025"
    ],
    "agentNotes": [
      "Ask a Collibra administrator to register an OAuth application. Integration applications use client credentials, and MCP clients need a user application with a redirect URI",
      "Reuse each access token for its 5 minutes. The token endpoint answers 429 once a client drains its pool of 100 tokens, refilled at 30 a minute",
      "Page large reads with `cursor=` and `sortField=ID`. Cursor and offset parameters cannot be combined, and Search API `limit` plus `offset` cannot exceed 10,000",
      "Keep Import API files to 10,000 assets or fewer and run at most 8 imports at once. Retry a job that fails on a concurrency error",
      "Expect `discover_data_assets` and `discover_business_glossary` to consume Collibra Units and to need the `dgc.ai-copilot` permission. Use `search_asset_keyword` for free keyword search"
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-10",
    "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"
  }
}