Terraform MCP Server by HashiCorp

# Fix list: Terraform MCP Server

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/terraform-mcp, the October 2026 research run, assessed 1 October 2026. Grade BB, 72.1 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 Terraform 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. Security & auth, 65 out of 100, up to 6.1 more on the total

Why it scored 65: Registry tools need no credentials. HCP Terraform tools take a user, team or organisation API token, which HCP Terraform can revoke, from `TFE_TOKEN` or a per-request `Authorization: Bearer` header in HTTP mode, and the server refuses a token sent as a query parameter with a 400. No OAuth (22 of 30). Delete, force-unlock, `action_run` and apply-capable runs stay off until `ENABLE_TF_OPERATIONS=true`, the default toolset is read-only, and HTTP deployments can pin an organisation allowlist. Creating and updating workspaces, setting variables, adding team members and granting team access still run without that flag once the terraform toolset is on (15 of 20). Module READMEs and provider docs from the public registry and run logs reach the model as they are. The README says not to use the server with untrusted clients or models, and HashiCorp's MCP security page names prompt injection and tool poisoning, without concrete mitigations (6 of 15). Tool-call logging middleware, opt-in OpenTelemetry metrics and a custom User-Agent so calls are attributable in HCP Terraform (10 of 15). SECURITY.md and a vulnerability policy with security@hashicorp.com, HashiCorp is a CVE Numbering Authority, security.txt valid per the 26 September check, no bug bounty, and the July 2026 security fixes went into the changelog with no advisory (12 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.

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

Why it scored 60: Free under MPL-2.0 with nothing to buy for the server, so 20 + 20 + 20. The registry half works with no account, which is the autonomous route. HCP Terraform usage is billed by HashiCorp on its own plans, which we didn't check. 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. Agent ergonomics, 74 out of 100, up to 4.2 more on the total

Why it scored 74: 61 tools in three toolsets, and only the nine registry tools load by default. `--toolsets` and `--tools` narrow the rest (22 of 25). List tools take `page` and `pageSize` (default 30) and search filters, but provider and module docs come back whole, with no section fetch (15 of 20). Errors are specific and some name the next step, for example a locked workspace points at `force_unlock_workspace` (15 of 20). readOnlyHint and destructiveHint are set on most tools, but the nine variable and variable-set tools carry no annotations, including `delete_variable_in_variable_set`, and there are no idempotency keys (12 of 20). Defaults for run type, page size and message, but `search_providers` needs four required arguments. go-tfe is the only official client library (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.

## 4. Reliability, 83 out of 100, up to 3.4 more on the total

Why it scored 83: Scored as a local package, since it runs as a Docker image, a Go binary or a self-hosted Streamable HTTP service. Official image hashicorp/terraform-mcp-server, binaries on releases.hashicorp.com and `go install`, with the Go version pinned in go.mod (20). Unit, end-to-end, HCP Terraform and security-scan workflows run on push to main. We couldn't see whether the default branch passes, so 20 of 25. 15 open issues, two bug reports open since January 2026 and several feature requests left open after the feature shipped. 0.5.1 was a hotfix for a race-condition crash in the MCP library (18 of 25). CHANGELOG.md calls out breaking changes in plain words, but 1.1.0 shipped two breaking changes in a minor version (10 of 15). Version 1.3.0 (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.

## 5. Schema & documentation, 80 out of 100, up to 3.3 more on the total

Why it scored 80: Every tool declares typed inputs through mcp-go, with required flags and enums (25). The README and the server's instructions.md are Markdown. We didn't find an llms.txt for the MCP docs and didn't check developer.hashicorp.com for one (5 of 10). The nine registry tools say when to call them, `search_providers` says it must run before `get_provider_details`, but most HCP Terraform tools are one line, such as 'Create a new variable in a variable set.' (14 of 20). Enums on `provider_document_type`, `run_type` and access levels, page sizes with defaults, a few free-form strings on workspace updates (12 of 15). Registry tools return a recovery hint on failure and the README has examples, but there's no error catalogue (9 of 15). Tagged releases and a CHANGELOG with dated GitHub tags (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. Maintenance & community, 78 out of 100, up to 1.9 more on the total

Why it scored 78: v1.3.0 tagged on 2026-08-25, 37 days before the run date, and 1.3.1 sits unreleased in the changelog (20). v1.1.0 (14 July), v1.2.0 (4 August) and v1.3.0 (25 August) inside 90 days (20). Commits land most days, including an SDK migration in late September, but the open issues include requests already shipped and bugs from January with no visible close-out (15 of 25). io.github.hashicorp/terraform-mcp-server is in the official registry under a GitHub-verified namespace, though the entry marked latest is 1.0.0 (13 of 15). CI, Dependabot bumps and a security-scan workflow (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, 89 out of 100, up to 1 more on the total

Made of editorial 78, provenance 100.

Why it scored 89: MPL-2.0 (30). Local software that talks to registry.terraform.io and the configured HCP Terraform or Terraform Enterprise address. Requests carry a User-Agent that identifies the MCP server to HashiCorp, which the changelog discloses. There's no MCP-specific data statement beyond HashiCorp's privacy policy (20 of 30). Breaking changes are flagged in the changelog when they ship. No deprecation policy or advance notice found (8 of 20). No telemetry by default. OpenTelemetry and Instana are opt-in and send to your own collector (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.

- -3: v1.1.0 (2026-07-14) fixed cross-tenant HCP Terraform token reuse in stateless HTTP mode and a client-supplied TFE_ADDRESS that could redirect the server's bearer token to any endpoint. Fixed and described in the changelog, but no advisory was published, so a reduced deduction (https://github.com/hashicorp/terraform-mcp-server/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.

- Whether the default branch CI passes; the run status wasn't visible to us.
- Whether developer.hashicorp.com publishes an llms.txt covering the MCP docs.
- Reply times on issues and pull requests.
- HCP Terraform pricing and free tier, which we didn't fetch.

## Weaknesses

- Workspace updates, variable writes, team membership and access grants run without the operations flag
- Nine variable and variable-set tools have no readOnly or destructive hints
- Provider docs come back whole, with no section-level fetch
- 1.1.0 fixed a cross-tenant token leak in HTTP mode with no published advisory
- The official registry entry still lists 1.0.0 as latest

## 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.

- Call `search_providers` first to get a `provider_doc_id`, then `get_provider_details` for that one document
- Pass the token in the `Authorization` header in HTTP mode. A `TFE_TOKEN` query parameter gets a 400
- Run `whoami` or `get_token_permissions` before writes to see what the token can do
- Check the workspace lock before `create_run`. A locked workspace returns an error naming `force_unlock_workspace`
- Pin the image tag (`1.3.0`), not `latest`, in scheduled jobs

## What the review panel asked for

- written deprecation policy
- registry entry kept current
- gate access grants
- advisories for fixes

## 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.