# Fix list: Azure DevOps MCP Server From Anchor Terminal's listing at https://www.anchorterminal.com/tools/azure-devops-mcp, the October 2026 research run, assessed 8 October 2026. Grade B, 66.3 out of 100. This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public. For a coding agent working on Azure DevOps 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. ## 1. Payments & pricing, 40 out of 100, up to 7.5 more on the total Why it scored 40: No x402, MPP or L402 (0). The server is free and Azure DevOps prices are public, with Basic at $6 a user a month after five free users (20). The first five users are free and the pricing page states no card requirement for them (20). A person has to create the organisation and sign in through Entra, and outside Microsoft's clients an administrator has to register an app (0). The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments): The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/). - 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which. - 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login. - 20, a free tier or trial that doesn't need a card. - 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API). Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied. Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol. ## 2. Schema & documentation, 64 out of 100, up to 5.9 more on the total Why it scored 64: Every tool in the local package has a zod schema turned into JSON Schema. The remote server's own definitions sit behind Entra sign-in and weren't read, and the docs list its tools and actions only (20). learn.microsoft.com has no llms.txt (404) but returns Markdown when asked with `Accept: text/markdown` (10). Descriptions are short, with a median of 89 characters across 45 tools, and most say only to use the `action` parameter. A few, such as `repo_pull_request_org`, say when to choose them (8). Enums, bounds and required organisation fields are present and no tool takes a free-form object. The dispatcher design leaves most parameters optional with a note such as "Used for update", so the schema can't mark what each action needs (10). TOOLSET.md lists parameters per tool and EXAMPLES.md gives prompts. Errors are plain-text tool results, and the troubleshooting page covers Entra error codes (8). The local package has semver releases with notes. The remote server is updated separately with no changelog found, and its tool page warns that the list may lag (8). The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema): APIs and MCP servers. - 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool). - 10, llms.txt or Markdown docs served for agents. - 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference. - 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs. - 0 to 15, examples and documented error responses. - 15, versioning and a public changelog. Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference. ## 3. Reliability, 72 out of 100, up to 5.6 more on the total Why it scored 72: Scored with the hosted lines, because the remote endpoint is the one Microsoft recommends. status.dev.azure.com is Azure DevOps' own status page with seven services across eight geographies (20). Its health API reported every service healthy on 8 October 2026 and the history page listed no events, but the page's filter is drawn by script and we couldn't confirm the view covers 90 days. The MCP endpoint is not a named component (15). Azure DevOps publishes a limit of 200 TSTUs per user in a sliding five-minute window, with delays of up to 30 seconds before blocking (15). Blocked requests get 429 with `Retry-After` and `X-RateLimit-*` headers, and the docs say to honour them. No retry guidance for writes was found (12). The data protection page states a 99.9 per cent SLA with refunds for Azure DevOps. Whether it covers the preview endpoint wasn't established (10). The remote server was announced as a public preview on 17 March 2026 and no general availability notice was found (0). The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability): Hosted APIs, MCP servers, models and platforms. - 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own). - 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so. - 15, rate limits documented with numbers. - 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved. - 10, an SLA published for any paid tier. - 10, the surface agents use is generally available, not beta or preview. Local packages, SDKs, frameworks and stdio MCP servers. - 20, installs from an official package with supported runtimes stated. - 25, a public CI and test suite, passing on the default branch. - 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered). - 15, semver discipline and breaking changes called out in a changelog. - 15, version 1.0 or later, or declared stable. Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors. ## 4. Agent ergonomics, 70 out of 100, up to 4.9 more on the total Why it scored 70: 48 tools documented for the remote server (5), with `X-MCP-Toolsets`, `X-MCP-Tools` and `X-MCP-Readonly` adding back 10. `top`, `skip` and continuation tokens on list and search tools, and the source trims pull request threads and branch lists before returning them (16). Errors name the missing parameter or pass Azure DevOps' message through as an `isError` result. Local sign-in failures in headless environments surface as a generic `fetch failed` per the troubleshooting guide (13). All 45 local tools set `readOnlyHint`, `destructiveHint`, `idempotentHint` and `openWorldHint`, with a test that fails on a missing entry, and work item updates accept a revision test for concurrency. There are no idempotency keys (16). The organisation comes from the URL and project and team defaults can be set by environment variable, but most calls still need a project. We confirmed the Node client library the server is built on, and didn't confirm a second official SDK (10). The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics): - 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries). - 20, pagination, filtering and output-size controls. - 20, actionable, documented error responses, codes and messages an agent can recover from. - 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations. - 15, sensible defaults, few required parameters, and official SDKs in at least two languages. Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs. ## 5. Security & auth, 81 out of 100, up to 3.3 more on the total Why it scored 81: The remote server accepts only Microsoft Entra OAuth with delegated Azure DevOps scopes, in the `Authorization` header, and a call needs both the app scope and the user's own permission. Conditional Access applies. The local package also takes PATs, which can be scoped and revoked (30). `X-MCP-Readonly`, toolset and per-tool headers, and guidance to grant read scopes only. No confirmation step for write tools was found, and `repo_pull_request_write` can set policy bypass on autocomplete when given a reason (14). The local source wraps returned content in random delimiters marked untrusted, with tests, though the Learn docs don't describe this or say whether the remote server does the same (11). Azure DevOps auditing exists for Entra-backed organisations and is itself labelled preview. Its event list has no MCP-specific events (8). SECURITY.md routes reports to MSRC, Azure DevOps is in Microsoft's bounty programme, and the data protection page lists SOC 2 Type 2 and ISO 27001:2022. The repository has no published advisories and microsoft.com's security.txt expired on 23 September 2026 (18). The checklist (https://www.anchorterminal.com/benchmark/#checklist-security): - 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option. - 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions. - 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10. - 0 to 15, audit logs or per-call visibility for the operator. - 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public. Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing. ## 6. Maintenance & community, 80 out of 100, up to 1.8 more on the total Why it scored 80: 2.10.0 was published to npm on 9 September 2026, 29 days before the check (30). Three stable releases in the last 90 days (2.8.1 on 14 July, 2.9.0 on 29 July, 2.10.0 on 9 September), nightly builds on the `next` tag, and commits as recent as 7 October (20). The issues page listed four open issues on 8 October, the oldest opened on 23 September, against issue numbers above 1,680. We couldn't read reply times (20). The official MCP registry has no entry under a Microsoft namespace. The repository's server.json names `com.microsoft/azure-devops`, but the registry returns only a third party's listing for this repository (0). Dependabot runs daily, and build, test, tool-name validation, CodeQL and dependency review workflows run on pull requests (10). The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance): - 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older. - 20, at least three releases or dated changelog entries in the last 90 days. - 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15. - 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models). - 10, package health, current dependencies and CI. Models are read for deprecation notice periods and model churn rather than release counts. ## 7. Transparency & trust, 79 out of 100, up to 1.8 more on the total Made of editorial 72, provenance 86. Why it scored 79: The local server is MIT, and the hosted service runs under Microsoft's online service terms (30). Azure DevOps documents encryption, replication, a 28-day recovery window for deleted organisations and its data locations. Nothing says what the remote MCP server logs or keeps, and the overview's privacy section says project information never leaves the customer's network, which doesn't fit a hosted endpoint (18). The README says the remote server will eventually replace the local one and asks users to plan migration, with no date. Dated notices exist for global PATs (1 December 2026) and Azure DevOps OAuth (10). Eight data geographies and the remote server's two outbound IP addresses are published. We didn't read a subprocessor list. The local source has no telemetry beyond a user agent string naming the MCP client (14). The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency): - 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms. - 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors). - 0 to 20, a deprecation policy or notices with dates. - 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted). The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two. Provenance checks not met in full (half of this category, computed from checked facts): - Terms of service: read, states 4 of the 7 things a reader expects, and has 2 clauses that cost points (3.4 of 10) - Privacy policy: read, states 8 of the 8 things a reader expects, and has 1 clause that costs points (8 of 10) - security.txt: published but past its Expires date (5 of 10) ## Deductions Each comes off the total. A fixed and documented problem counts for less at the next check. - 2026-07-29: version 2.9.0 of the local package renamed its tools in a minor release while consolidating them to match the remote server. The README warning and the advice to pin 2.8.1 were added on 31 July, and no aliases were kept. Documented since, so the deduction is small (https://github.com/microsoft/azure-devops-mcp/pull/1476). ## What we couldn't check What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it. - unchecked: whether the remote server reached general availability. The roadmap lists it under 2026 Q3, and we found only the preview announcement and a README that still mentions the preview period - unchecked: the 90-day incident record. The status history page listed no events but its date filter is drawn by script - unchecked: the remote server's own tool definitions and annotations, which need an Entra sign-in to list - unchecked: the Azure DevOps entry in the Product Terms and the SLA text, both on pages drawn by script, and whether the SLA covers a preview endpoint - unchecked: reply times on GitHub issues and the exact star count (the page shows 2.0k) - unchecked: whether creating an Azure DevOps organisation asks for a payment card - What the remote server logs and how long it keeps it, which no page we read states - The overview page (12 August 2026) says Claude Code and Cursor must use the local server, while the remote page (21 September 2026) gives them a route through a custom Entra app registration ## Weaknesses - The remote server was announced as a public preview on 17 March 2026 and no general availability notice was found - Claude Code and Cursor need a custom Entra app registration with admin consent, and Claude Desktop and Codex can't use the remote server - Version 2.9.0 renamed the local tools in a minor release on 29 July 2026, with the README warning added two days later - No entry under a Microsoft namespace in the official MCP registry. The only listing naming this repository is a third party's - Most tool descriptions are one line ending in an instruction to use the `action` parameter, with a median of 89 characters ## What costs an agent a turn today The notes we give agents before they call it. Each one is a workaround an agent shouldn't need. - Put the organisation in the URL (`https://mcp.dev.azure.com/{organization}`). Without it every call must name the organisation, and guest users can't connect at all - Send `X-MCP-Readonly: true` for read tasks and `X-MCP-Toolsets: repos,wit` to cut the tool list. Don't combine `X-MCP-Toolsets` with `X-MCP-Tools` - Name the project in each call. A missing project is the error the vendor reports most often - For headless runs use the local package with `--authentication envvar` and a token in `ADO_MCP_AUTH_TOKEN`. The default interactive sign-in fails without a browser - On HTTP 429 wait for `Retry-After`. Azure DevOps delays or blocks a user above 200 TSTUs in a sliding five-minute window ## When it's done Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.