HashiCorp Vault + Vault MCP Server by HashiCorp (IBM)

# Fix list: HashiCorp Vault + Vault MCP Server

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/hashicorp-vault, the October 2026 research run, assessed 1 October 2026. Grade B, 64.4 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 HashiCorp Vault + Vault 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, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 (0). HCP Vault Dedicated hourly prices per cluster and $72.92 a client a month are public on the IBM price table, marked indicative (20). A $500 HCP trial credit, and the pricing pages don't say whether a card is needed (10 of 20). A person signs up for HCP and creates the cluster, and there's no programmatic signup (0). Vault Community self-hosts free under the BUSL-1.1, but the rubric scores the paid option.

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. Agent ergonomics, 64 out of 100, up to 5.9 more on the total

Why it scored 64: The MCP server has 16 tools with no toolsets or read-only subset (15 of 25, KV and LIST responses are small). LIST returns keys only and KV v2 reads one version, but there's little paging or filtering on most LIST endpoints (10 of 20). Errors come back as an errors array of strings with an HTTP status, which says permission denied without naming the missing capability (12 of 20). Every MCP tool carries readOnlyHint or destructiveHint, and KV v2 writes take a cas version for safe retries (17 of 20). The official client is the Go api package; Python (hvac) and Node clients are community. Vault Agent and Vault Proxy handle auto-auth, renewal and caching so an app reads a file or a local socket (10 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.

## 3. Reliability, 71 out of 100, up to 5.8 more on the total

Why it scored 71: Status page at status.hashicorp.com (incident.io) with history for HCP products, including 23 Vault Dedicated region components per the 30 September check (20). Between 1 July and 1 October 2026 nothing was posted against HCP Vault Dedicated. The incidents were on HCP Terraform, Infragraph and Vault Radar, plus a high-severity DR cluster creation failure on HCP on 6 August whose product isn't named and a releases.hashicorp.com outage on 26 September that blocks binary downloads, so we count minor only with a doubt (25 of 30). Rate limit quotas are documented and configurable, but no default figures are published for HCP or a fresh install (7 of 15). Quotas can send Retry-After and X-Ratelimit headers when enable_rate_limit_response_headers is set, and KV v2 has check-and-set for safe writes (12 of 15). No SLA found on the HashiCorp or IBM pricing pages (0). The HTTP API is generally available, while the MCP server is beta (7 of 10).

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. Schema & documentation, 74 out of 100, up to 4.2 more on the total

Why it scored 74: Every Vault server generates an OpenAPI document at /v1/sys/internal/specs/openapi from its mounted paths, so the contract matches the plugins you run, though it isn't a static file you can read before installing (22 of 25). No llms.txt at developer.hashicorp.com or under /vault (both 404) (0). API pages state purpose per endpoint, and the MCP tools say when to be careful (delete_mount says "Use with extreme caution"), but none say when not to use them (14 of 20). Typed parameters in the OpenAPI and typed, required arguments on the 16 MCP tools (12 of 15). Sample requests and responses on each API page and a documented errors array with status codes (13 of 15). /v1 is stable and CHANGELOG.md dates every release with BREAKING CHANGES sections, but the MCP server's changelog lists a 0.2.1 that was never released (13 of 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.

## 5. Security & auth, 86 out of 100, up to 2.5 more on the total

Why it scored 86: Tokens carry TTLs and path policies, machines log in with AppRole, Kubernetes, JWT/OIDC, AWS, GCP, Azure or TLS certificates, dynamic secrets get leases that revoke on expiry, and Enterprise 2.x adds an OAuth resource server so an agent can present a JWT instead of a Vault token (30). Path ACLs with explicit deny and read-only capabilities, control groups for approvals on Enterprise, and agent ceiling policies on Enterprise; the MCP server has no read-only mode (17 of 20). Secrets aren't untrusted content, but the MCP README warns that read_secret puts values in front of the model (10). Audit devices (file, syslog, socket) on every edition, with HMAC'd values (15). security.txt with a contact and policy but no Expires field, CVEs named in the changelog for every patch release, but two security fixes in the MCP server (28 July and 11 August 2026) sit unreleased with no advisory (14 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.

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

Why it scored 77: Vault 2.1.1 on 16 September 2026, 15 days before this check (30). 2.0.4 (4 August), 2.1.0 (1 September) and 2.1.1 (16 September), plus 1.21.x Enterprise patches (20). 1.2k open issues, labelled for triage, including open regressions from 29 July (#32059) and 5 August (#32072), and the newest open issue we saw was from 10 August (12 of 25). Only the Go client is official, and the MCP server isn't in the MCP registry, where HashiCorp lists only the Terraform server (8 of 15). CI runs Go tests, security scans and Enos scenarios, but the MCP server's newest binary and Docker image are 0.2.0 from 24 September 2025 (7 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.

## 7. Transparency & trust, 83 out of 100, up to 1.5 more on the total

Made of editorial 65, provenance 100.

Why it scored 83: Vault is BUSL-1.1 with IBM as licensor, converting to MPL 2.0 four years after each version, and the MCP server is MPL-2.0. Source is public with clear terms, but the BUSL isn't an OSI licence (18 of 30). Privacy policy of 20 April 2026 per the 30 September check, and self-hosted Community sends nothing to HashiCorp; we didn't read an HCP DPA or subprocessor list (15 of 30). Dated end-of-sale and end-of-life notices for HCP Vault Secrets and BREAKING CHANGES sections in the changelog (16 of 20). The Community build stubs out the census reporter, so nothing phones home; Enterprise licence reporting wasn't checked (16 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.

## Deductions

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

- -4: The official Vault MCP server fixed cross-user credential inheritance through a shared MCP session ID on 2026-07-28 and an SSRF through a VAULT_ADDR query parameter on 2026-08-11, but the newest binary and Docker image are still 0.2.0 from 2025-09-24 and no advisory was published (https://github.com/hashicorp/vault-mcp-server/commits/main, https://releases.hashicorp.com/vault-mcp-server/)
- -1: Vault 2.0.3 (2026-06-17) fixed a LIST ACL bypass where a trailing slash skipped a more specific deny rule; fixed and documented in the changelog, so it decays (https://github.com/hashicorp/vault/blob/main/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.

- The listing gave the MCP server as version 0.2.1; the newest build on releases.hashicorp.com and Docker Hub is 0.2.0 from 24 September 2025, corrected in the patch.
- We dropped the listing's -3 for the HCP Vault Secrets retirement, since it came with a dated year of notice and fits none of the negative categories.
- Whether HCP Vault Dedicated carries an SLA; none appeared on the HashiCorp or IBM pricing pages.
- unchecked: an HCP DPA and subprocessor list, and whether the $500 HCP credit needs a card.
- Which product the 6 August DR cluster creation failure on status.hashicorp.com affected.

## Weaknesses

- The MCP server's newest build is 0.2.0 from September 2025, and security fixes from July and August 2026 are unreleased
- Agentic IAM, control groups and the OAuth resource server are Enterprise only
- BUSL-1.1, not an OSI licence, and HCP Vault Secrets was retired within two years of launch
- No llms.txt, and the only official client library is Go
- HCP client pricing ($72.92 a client a month) can dwarf the cluster price for many agents, and no SLA is published

## 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.

- Prefer a dynamic secret (database, AWS, GCP engines) over a KV read; the lease expires with the run and revoke is one call
- Log in with AppRole or Kubernetes auth and keep the token for its TTL. Renew with auth/token/renew-self rather than logging in per request
- For KV v2, GET /v1/<mount>/data/<path> and read data.data, and pass cas on writes so a retry can't overwrite a newer version
- If you must use the MCP server, build it from main rather than running the 0.2.0 image, run it over stdio, and give it a token limited to one mount
- Ask your operator to set enable_rate_limit_response_headers on the quota so a 429 carries Retry-After

## What the review panel asked for

- a current MCP release
- release the MCP fixes
- read-only MCP mode

## 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.