Azure MCP Server by Microsoft

# Fix list: Azure MCP Server

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/azure-mcp, the October 2026 research run, assessed 1 October 2026. Grade B, 67.8 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 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. Reliability, 69 out of 100, up to 6.2 more on the total

Why it scored 69: Scored as a local package, though it can also be self-hosted over HTTP. Official packages on npm (Node 22 or later stated), NuGet, PyPI and MCPB (20). Tests and live-test pipelines exist, but CI runs outside GitHub Actions and we couldn't see whether the default branch passes (15 of 25). 256 open issues across the repository, each carrying triage labels, with recent bug reports on proxy bypass and Container Apps connections (15 of 25). Every release in CHANGELOG.md has a Breaking Changes section, but breaking changes land twice a week inside 3.0.0 prereleases (12 of 15). 2.0.2 (24 April 2026) is the stable line, yet the npm `latest` tag points at 3.0.0-beta.49, so `npx @azure/mcp@latest` installs a beta (7 of 15).

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.

## 2. Payments & pricing, 60 out of 100, up to 5 more on the total

Why it scored 60: Free under MIT with nothing to buy for the server, so 20 + 20 + 20. Every tool acts on an Azure subscription that Microsoft bills separately, which this grade doesn't cover. No payment protocol (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.

## 3. Security & auth, 73 out of 100, up to 4.7 more on the total

Why it scored 73: Entra ID through DefaultAzureCredential (az login, managed identity, service principal, workload identity federation), so access is RBAC-scoped and nothing secret sits in the MCP config. HTTP mode authenticates incoming callers and can act on behalf of the user (28 of 30). `--read-only` drops write tools, `--namespace` narrows the surface, and reading secrets, connection strings or private keys asks the user first through elicitation. Destructive operations get no confirmation, which the README says plainly (14 of 20). Monitor and Application Insights queries, blobs and database rows reach the model as they are. We found no prompt-injection guidance, though SSRF and query-injection hardening went in this year (5 of 15). Azure's Activity Log records the writes made with the caller's identity, and telemetry can go to your own Application Insights (11 of 15). SECURITY.md sends reports to MSRC, Microsoft runs a bug bounty and MSRC assigned CVEs this year, but no GitHub advisories are published and microsoft.com's security.txt expired on 23 September 2026 per the 26 September check (15 of 20).

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.

## 4. Agent ergonomics, 72 out of 100, up to 4.6 more on the total

Why it scored 72: About 430 commands across roughly 60 service namespaces. The default namespace mode shows one tool per namespace, still over 30, but `--namespace`, `--tool`, consolidated mode and a single-tool mode shrink it (15 of 25). Resource Graph queries, pagination on some data tools and a compact structured-output mode (14 of 20). Errors are specific in the tools that have been reworked, and the router lists valid command names instead of dumping help (15 of 20). Every command carries destructive, idempotent, readOnly, openWorld and secret metadata, and Destructive defaults to true when unset. The email and SMS send tools were marked read-only until 1 October 2026 (16 of 20). DefaultAzureCredential needs no config, packages exist for Node, .NET and Python, but subscription and resource group are required on most tools (12 of 15).

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. Schema & documentation, 77 out of 100, up to 3.7 more on the total

Why it scored 77: Every command has a typed JSON Schema, and 3.0.0-beta.48 fixed the CLI metadata that reported every option as a string. In the default namespace mode each tool takes a `command` name and a nested parameter object (22 of 25). docs/azmcp-commands.md and llms-install.md are Markdown in the repository; we didn't find an llms.txt (5 of 10). Descriptions state purpose and the router returns the available command names on a miss, but they assume Azure vocabulary and rarely say when not to call (13 of 20). Typed options and enums per command, flattened into a routing object in namespace mode (10 of 15). Usage examples for about 430 commands and e2e test prompts. Storage errors now map 403, 404 and 409 to specific messages (12 of 15). A detailed CHANGELOG with dated releases (15).

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.

## 6. Transparency & trust, 77 out of 100, up to 2 more on the total

Made of editorial 71, provenance 82.

Why it scored 77: MIT (30). Local software that calls Azure APIs with your identity. Telemetry to Microsoft is described in the README under Microsoft's privacy statement, with no MCP-specific retention statement (18 of 30). Breaking changes are listed per release, but tools are removed or renamed between prereleases with no notice period, such as `resilience_*` becoming `resiliency_*` on 22 September and the ADME tools pulled on 1 October (8 of 20). Telemetry to Microsoft is on by default, documented, with `AZURE_MCP_COLLECT_TELEMETRY=false` and a Microsoft-only opt-out (15 of 20).

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):

- Status page: not found (0 of 10)
- security.txt: published but past its Expires date (5 of 10)

## 7. Maintenance & community, 89 out of 100, up to 1 more on the total

Why it scored 89: 3.0.0-beta.49 on 2026-10-01 (30). 26 releases between 8 July and 1 October, on a stated Tuesday and Thursday cadence (20). Issues get area and triage labels quickly and fixes reference them, but several fixed issues stay open and we couldn't see reply times (16 of 25). com.microsoft/azure is in the official registry under a DNS-verified namespace (15). Dependency and .NET SDK updates ship with releases. Build status wasn't visible (8 of 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.

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- -2: CVE-2026-26118, published 2026-03-10, CVSS 8.8. Server-side request forgery in Azure MCP Server let an authorised attacker elevate privileges over a network. Fixed and published through MSRC (https://nvd.nist.gov/vuln/detail/CVE-2026-26118).
- -2: CVE-2026-32211, published 2026-04-03, CVSS 9.1. Missing authentication for a critical function in Azure MCP Server let an unauthorised attacker disclose information over a network. Fixed and published through MSRC (https://nvd.nist.gov/vuln/detail/CVE-2026-32211).
- -1: until 3.0.0-beta.49 on 2026-10-01, `communication_email_send` and `communication_sms_send` were annotated read-only, so they stayed available under `--read-only`, an outbound send path in a mode meant to block writes. Fixed and described in the changelog (https://github.com/microsoft/mcp/blob/main/servers/Azure.Mcp.Server/CHANGELOG.md).

## 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.

- Whether CI passes on the default branch; builds run in Azure Pipelines that we didn't read.
- Which versions CVE-2026-26118 and CVE-2026-32211 affected and which release fixed them; the NVD summaries don't say.
- Whether CVE-2026-33980 (KQL injection in an 'Azure Data Explorer MCP Server') concerns this server's Kusto tools, which got KQL injection fixes in March 2026, or a separate project.
- When 3.0.0 reaches a stable release.

## Weaknesses

- npm `latest` installs a 3.0.0 beta, and betas rename and remove tools without a notice period
- No confirmation step before deletes and other destructive calls
- Telemetry to Microsoft is on by default
- Two MSRC CVEs in March and April 2026, rated 8.8 and 9.1
- Default namespace mode still exposes about 60 tools

## 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.

- Start with `--namespace <one or two> --read-only` for inspection and widen only when a task needs a write
- Pin a version instead of `@latest`, which is a prerelease
- Use `resiliency_*`, not `resilience_*`. The prefix changed on 22 September 2026
- Resolve the subscription and resource group once and pass them on every call
- Auth failures mean the credential chain found nothing. Run `az login` or set the service-principal variables

## What the review panel asked for

- latest tag on the stable line
- notice before renames
- confirmation on deletes
- affected versions published

## 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.