{
  "fixes": {
    "slug": "functionize-mcp",
    "name": "Functionize MCP server",
    "listing": "https://www.anchorterminal.com/tools/functionize-mcp",
    "markdown": "# Fix list: Functionize MCP server\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/functionize-mcp, the October 2026 research run, assessed 24 October 2026. Grade E, 44.7 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 Functionize MCP server: 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. Schema \u0026 documentation, 29 out of 100, up to 11.5 more on the total\n\nWhy it scored 29: No machine-readable contract could be read. `tools/list` answers 401 without a signed-in account, and no public tool manifest, server card or source was found, so typed inputs on the tools are unchecked (0 of 25). llms.txt is an August 2025 crawl of the marketing site with no mention of MCP, but SETUP.md is Markdown written for an agent to run and the two skills are Markdown references for agents (10). Purpose and when to use each tool come only from the skills, which describe the session loop and say what Studio can't do; the tool descriptions themselves weren't read (8 of 20). Typed inputs, enums and constraints unread (0 of 15). Example prompts in the release posts and skills, and a troubleshooting table in SETUP.md with the 403 on a wrong team, `redirect_uri not permitted` and the unknown-tool reply; no documented error shape for tool calls (8 of 15). The plugin is versioned (1.1.0, skills 2.12.0 and 2.40.0) with a dated git history, but the server has no version or changelog (3 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## 2. Reliability, 50 out of 100, up to 10 more on the total\n\nWhy it scored 50: Graded with the hosted lines, on the remote MCP server. Functionize has a status page at status.functionize.com (incident.io) with incident history, but its components are app.functionize.com and two virtualautomationengineer.com hosts, not Studio or mcp.functionize.com. Judgement call, 10 of 20 for a page that doesn't name the surface graded. From 26 July to 24 October 2026 it records one minor incident, slower test creation and execution for 11 minutes on 22 October; the major incident of 23 July falls just outside the window (20). No request rate limit is published with numbers; the only limits are 5 or 10 parallel runs and monthly credits per plan (5 of 15). No 429 or Retry-After guidance was found. The plugin's skill says a create submitted before a dropped connection finishes on the server and to list recent tests before creating again, which is safe-retry guidance of a kind (5 of 15). The terms say self-service plans have no SLA and service levels apply only where an Order Form states them, with no figure published (0). The release of 8 October calls the server available today and nothing labels it beta (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. Agent ergonomics, 43 out of 100, up to 9.3 more on the total\n\nWhy it scored 43: Graded on the MCP server. 10 tools by the README's count, with the long guidance moved into Claude Code skills that load only when used. The size of the definitions couldn't be read, so 20 of 25 rather than 25 for compact tools. Sessions can be listed and their events read or streamed, but no page size, filter or output-size control was found (5 of 20). Results carry the failing step, screenshots and logs, and SETUP.md maps symptoms to fixes; no error codes for tool calls were found (8 of 20). No idempotency key, and readOnlyHint and destructiveHint couldn't be checked; the skill's advice to list recent tests before creating again is the only retry guidance (3 of 20). Requests are plain words, the team defaults to the account's own and sign-in needs no key, but there's no official SDK (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## 4. Payments \u0026 pricing, 30 out of 100, up to 8.8 more on the total\n\nWhy it scored 30: No x402, MPP or L402 (0). Plan prices are public ($0, $20 and $100 a month with their credits, add-on credits at $10 per 200,000), but the credits a task uses aren't published, so plan-only pricing (10). A Free plan with no card and no time limit (20). A person has to sign up for Studio and sign in through a browser before any call works (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## 5. Security \u0026 auth, 56 out of 100, up to 7.7 more on the total\n\nWhy it scored 56: OAuth only. The metadata gives the authorisation code grant with PKCE S256, refresh tokens, dynamic client registration, client ID metadata documents, public clients and a revocation endpoint, with scopes `agent.sessions:read`, `agent.sessions:write` and `offline_access`. No API key exists and tokens go in the header only (30). The connection's reach is the whole team. Per SETUP.md it can read, continue or cancel any session in the team, including colleagues', and delete attachments on one. No documented way to request the read scope alone, and no approval step for writes, though the server has no file-system access and Functionize says Studio doesn't decide releases (8 of 20). Results include text and screenshots from the app under test, which is the customer's own but untrusted. No guidance on that was found; SETUP.md's setup prompt confines the agent to two configuration changes and tells it to stop if asked for more, and the skills keep credentials in masked variables (5 of 15). Sessions and run history with screenshots and logs are visible in Studio, and usage in the sidebar and reports; an audit-logs row is commented out of the pricing page (7 of 15). No security.txt (404 on both hosts), no bug bounty or advisories found, security@functionize.com named in the terms, SOC 2 Type II adherence claimed in the privacy policy with no report public, and trust.functionize.com behind a login (6 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## 6. Maintenance \u0026 community, 49 out of 100, up to 4.5 more on the total\n\nWhy it scored 49: Read as a closed hosted service with a public plugin repository. The server was released on 8 October 2026, 16 days ago (30). The plugin repository has 19 dated commits from 26 August to 10 October 2026 and one version bump (1.1.0, 16 September), and the blog has the release of 8 October; no tagged releases or changelog, so 10 of 20. Support is through the in-app system with published response goals (2 hours for critical and high, 4 for medium, 24 for low), stated as goals, not commitments. No public changelog or community channel, and the GitHub issue pages answered 403 to us (6 of 25). Not in the official MCP registry (searches for functionize and com.functionize returned nothing on 24 October 2026) and no official SDK (0). The plugin is MIT with a current manifest but no CI or tests (3 of 10). The server is 16 days old by its release.\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, 59 out of 100, up to 3.6 more on the total\n\nMade of editorial 38, provenance 80.\n\nWhy it scored 59: The editorial half. A closed service under dated terms naming Functionize, Inc. and California law (version 2.0, effective 10 July 2026), with the plugin and skills under MIT (17 of 30). The privacy policy (effective 10 July 2026) and terms agree that customer content isn't used to train foundation models except when aggregated and de-identified, while usage data may be used for training at Functionize's discretion. Deletion or anonymisation follows termination within about sixty days, sandbox data may be reset, and a DPA is by agreement only. No retention period for run artifacts is given (15 of 30). No deprecation policy; SETUP.md mentions a retired server address with no dated notice (2 of 20). Stripe is the only named sub-processor; cloud hosting and AI model providers are unnamed and no data location is stated (4 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- Changelog: not found (0 of 10)\n- security.txt: not found (0 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- unchecked: the tools as served, because `initialize` and `tools/list` answered 401 without a signed-in account and no public manifest or source exists. The count of 10 is the plugin README's and the names come from the skills; schemas, descriptions and annotations are unread.\n- unchecked: whether the Free plan's account can use the MCP tools without a sales step. The release says customers and trial users, the pricing page has no MCP row (an \"API and MCP access\" row is commented out of its HTML), and the skill says agent-session access must be enabled for the team. Listed because a free account anyone can open is the documented way in.\n- firstReleased is 8 October 2026, the date shown on Functionize's release saying the server is \"available today\". That's earlier than the lead's date of about 23 October. The how-to post shows 8 October but its structured data says 19 October 2026. The plugin repository dates from 26 August 2026 and a /mcp-waitlist page is still published, so the server may have been open to some accounts earlier; no first-hand page shows that.\n- The Free plan's monthly credits are 800,000 on the plan card and 400,000 in the FAQ of the same page, beside a launch promotion of 2,000,000 bonus credits. Recorded, not deducted.\n- Category. browser, because Studio tests web apps by driving cloud browsers (Chrome, Firefox and Edge) and returns steps, screenshots and logs. The coding agent doesn't drive the browser itself; Studio's agent does, as with Skyvern. No testing category exists. code was the alternative. Judgement call.\n- Our own products. The Functionize MCP server doesn't do LocalGhost's job or letme's. It reaches one vendor's testing service and chooses no tools. Graded the normal way.\n\n## Weaknesses\n\n- Tool definitions, input schemas and annotations aren't published, and `tools/list` needs a signed-in account\n- No numeric rate limit, 429 guidance or idempotency key was found; only parallel-run limits of 5 or 10 per plan\n- The status page covers the legacy app hosts, not Studio or mcp.functionize.com\n- The connection's reach is the whole team, so it can read, continue or cancel colleagues' sessions\n- Credits per task aren't published, and the Free plan's monthly credits are given as 800,000 and 400,000 on the same page\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- Have the person complete the browser sign-in, which an agent can't do, then check the connection with `list_agent_teams`\n- Pass the project ID, environment and test ID in `start_agent_session`, and continue a waiting session with `send_agent_message` rather than starting another\n- Report a created test as done only after a green run, not on a test ID or a green generation\n- After a dropped connection, list the project's recent tests before creating a test again, since a submitted create finishes on the server\n- Send the `X-Functionize-Team-Id` header to pin a team, and keep credentials in masked Studio variables, never in prompts\n\n## What the review panel asked for\n\n- Publish a tool manifest\n- Add a server changelog\n- Publish tool definitions\n- State MCP access per plan\n- Add Studio and the MCP server to the status page\n- Publish request limits and 429 behaviour\n- a read-only connection option\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": "E",
    "score": 44.7,
    "assessed": "2026-10-24",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "score": 29,
        "maxGain": 11.5,
        "reason": "No machine-readable contract could be read. `tools/list` answers 401 without a signed-in account, and no public tool manifest, server card or source was found, so typed inputs on the tools are unchecked (0 of 25). llms.txt is an August 2025 crawl of the marketing site with no mention of MCP, but SETUP.md is Markdown written for an agent to run and the two skills are Markdown references for agents (10). Purpose and when to use each tool come only from the skills, which describe the session loop and say what Studio can't do; the tool descriptions themselves weren't read (8 of 20). Typed inputs, enums and constraints unread (0 of 15). Example prompts in the release posts and skills, and a troubleshooting table in SETUP.md with the 403 on a wrong team, `redirect_uri not permitted` and the unknown-tool reply; no documented error shape for tool calls (8 of 15). The plugin is versioned (1.1.0, skills 2.12.0 and 2.40.0) with a dated git history, but the server has no version or changelog (3 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": "reliability",
        "name": "Reliability",
        "score": 50,
        "maxGain": 10,
        "reason": "Graded with the hosted lines, on the remote MCP server. Functionize has a status page at status.functionize.com (incident.io) with incident history, but its components are app.functionize.com and two virtualautomationengineer.com hosts, not Studio or mcp.functionize.com. Judgement call, 10 of 20 for a page that doesn't name the surface graded. From 26 July to 24 October 2026 it records one minor incident, slower test creation and execution for 11 minutes on 22 October; the major incident of 23 July falls just outside the window (20). No request rate limit is published with numbers; the only limits are 5 or 10 parallel runs and monthly credits per plan (5 of 15). No 429 or Retry-After guidance was found. The plugin's skill says a create submitted before a dropped connection finishes on the server and to list recent tests before creating again, which is safe-retry guidance of a kind (5 of 15). The terms say self-service plans have no SLA and service levels apply only where an Order Form states them, with no figure published (0). The release of 8 October calls the server available today and nothing labels it beta (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": "ergonomics",
        "name": "Agent ergonomics",
        "score": 43,
        "maxGain": 9.3,
        "reason": "Graded on the MCP server. 10 tools by the README's count, with the long guidance moved into Claude Code skills that load only when used. The size of the definitions couldn't be read, so 20 of 25 rather than 25 for compact tools. Sessions can be listed and their events read or streamed, but no page size, filter or output-size control was found (5 of 20). Results carry the failing step, screenshots and logs, and SETUP.md maps symptoms to fixes; no error codes for tool calls were found (8 of 20). No idempotency key, and readOnlyHint and destructiveHint couldn't be checked; the skill's advice to list recent tests before creating again is the only retry guidance (3 of 20). Requests are plain words, the team defaults to the account's own and sign-in needs no key, but there's no official SDK (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": "payments",
        "name": "Payments \u0026 pricing",
        "score": 30,
        "maxGain": 8.8,
        "reason": "No x402, MPP or L402 (0). Plan prices are public ($0, $20 and $100 a month with their credits, add-on credits at $10 per 200,000), but the credits a task uses aren't published, so plan-only pricing (10). A Free plan with no card and no time limit (20). A person has to sign up for Studio and sign in through a browser before any call works (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": "security",
        "name": "Security \u0026 auth",
        "score": 56,
        "maxGain": 7.7,
        "reason": "OAuth only. The metadata gives the authorisation code grant with PKCE S256, refresh tokens, dynamic client registration, client ID metadata documents, public clients and a revocation endpoint, with scopes `agent.sessions:read`, `agent.sessions:write` and `offline_access`. No API key exists and tokens go in the header only (30). The connection's reach is the whole team. Per SETUP.md it can read, continue or cancel any session in the team, including colleagues', and delete attachments on one. No documented way to request the read scope alone, and no approval step for writes, though the server has no file-system access and Functionize says Studio doesn't decide releases (8 of 20). Results include text and screenshots from the app under test, which is the customer's own but untrusted. No guidance on that was found; SETUP.md's setup prompt confines the agent to two configuration changes and tells it to stop if asked for more, and the skills keep credentials in masked variables (5 of 15). Sessions and run history with screenshots and logs are visible in Studio, and usage in the sidebar and reports; an audit-logs row is commented out of the pricing page (7 of 15). No security.txt (404 on both hosts), no bug bounty or advisories found, security@functionize.com named in the terms, SOC 2 Type II adherence claimed in the privacy policy with no report public, and trust.functionize.com behind a login (6 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": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 49,
        "maxGain": 4.5,
        "reason": "Read as a closed hosted service with a public plugin repository. The server was released on 8 October 2026, 16 days ago (30). The plugin repository has 19 dated commits from 26 August to 10 October 2026 and one version bump (1.1.0, 16 September), and the blog has the release of 8 October; no tagged releases or changelog, so 10 of 20. Support is through the in-app system with published response goals (2 hours for critical and high, 4 for medium, 24 for low), stated as goals, not commitments. No public changelog or community channel, and the GitHub issue pages answered 403 to us (6 of 25). Not in the official MCP registry (searches for functionize and com.functionize returned nothing on 24 October 2026) and no official SDK (0). The plugin is MIT with a current manifest but no CI or tests (3 of 10). The server is 16 days old by its release.",
        "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": 59,
        "maxGain": 3.6,
        "reason": "The editorial half. A closed service under dated terms naming Functionize, Inc. and California law (version 2.0, effective 10 July 2026), with the plugin and skills under MIT (17 of 30). The privacy policy (effective 10 July 2026) and terms agree that customer content isn't used to train foundation models except when aggregated and de-identified, while usage data may be used for training at Functionize's discretion. Deletion or anonymisation follows termination within about sixty days, sandbox data may be reset, and a DPA is by agreement only. No retention period for run artifacts is given (15 of 30). No deprecation policy; SETUP.md mentions a retired server address with no dated notice (2 of 20). Stripe is the only named sub-processor; cloud hosting and AI model providers are unnamed and no data location is stated (4 of 20).",
        "blend": "editorial 38, provenance 80",
        "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": "Changelog",
        "value": "not found",
        "points": 0,
        "max": 10
      },
      {
        "label": "security.txt",
        "value": "not found",
        "points": 0,
        "max": 10
      }
    ],
    "unchecked": [
      "unchecked: the tools as served, because `initialize` and `tools/list` answered 401 without a signed-in account and no public manifest or source exists. The count of 10 is the plugin README's and the names come from the skills; schemas, descriptions and annotations are unread.",
      "unchecked: whether the Free plan's account can use the MCP tools without a sales step. The release says customers and trial users, the pricing page has no MCP row (an \"API and MCP access\" row is commented out of its HTML), and the skill says agent-session access must be enabled for the team. Listed because a free account anyone can open is the documented way in.",
      "firstReleased is 8 October 2026, the date shown on Functionize's release saying the server is \"available today\". That's earlier than the lead's date of about 23 October. The how-to post shows 8 October but its structured data says 19 October 2026. The plugin repository dates from 26 August 2026 and a /mcp-waitlist page is still published, so the server may have been open to some accounts earlier; no first-hand page shows that.",
      "The Free plan's monthly credits are 800,000 on the plan card and 400,000 in the FAQ of the same page, beside a launch promotion of 2,000,000 bonus credits. Recorded, not deducted.",
      "Category. browser, because Studio tests web apps by driving cloud browsers (Chrome, Firefox and Edge) and returns steps, screenshots and logs. The coding agent doesn't drive the browser itself; Studio's agent does, as with Skyvern. No testing category exists. code was the alternative. Judgement call.",
      "Our own products. The Functionize MCP server doesn't do LocalGhost's job or letme's. It reaches one vendor's testing service and chooses no tools. Graded the normal way."
    ],
    "weaknesses": [
      "Tool definitions, input schemas and annotations aren't published, and `tools/list` needs a signed-in account",
      "No numeric rate limit, 429 guidance or idempotency key was found; only parallel-run limits of 5 or 10 per plan",
      "The status page covers the legacy app hosts, not Studio or mcp.functionize.com",
      "The connection's reach is the whole team, so it can read, continue or cancel colleagues' sessions",
      "Credits per task aren't published, and the Free plan's monthly credits are given as 800,000 and 400,000 on the same page"
    ],
    "agentNotes": [
      "Have the person complete the browser sign-in, which an agent can't do, then check the connection with `list_agent_teams`",
      "Pass the project ID, environment and test ID in `start_agent_session`, and continue a waiting session with `send_agent_message` rather than starting another",
      "Report a created test as done only after a green run, not on a test ID or a green generation",
      "After a dropped connection, list the project's recent tests before creating a test again, since a submitted create finishes on the server",
      "Send the `X-Functionize-Team-Id` header to pin a team, and keep credentials in masked Studio variables, never in prompts"
    ],
    "requests": [
      {
        "text": "Publish a tool manifest",
        "reviews": 1
      },
      {
        "text": "Add a server changelog",
        "reviews": 1
      },
      {
        "text": "Publish tool definitions",
        "reviews": 1
      },
      {
        "text": "State MCP access per plan",
        "reviews": 1
      },
      {
        "text": "Add Studio and the MCP server to the status page",
        "reviews": 1
      },
      {
        "text": "Publish request limits and 429 behaviour",
        "reviews": 1
      },
      {
        "text": "a read-only connection option",
        "reviews": 1
      }
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-25",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.6",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  }
}