◎ Discry Score
hashicorp.com
infrastructure · API
B
0 / 100
DISCOVERY0
COMPREHENSION0
Category leader: 98 (A)
Discry your API →
INFRASTRUCTURE · RANK #33 OF 68

HashiCorp is Good to agents.

Discry independently scored how well an AI agent can discover and understand the HashiCorp API from what’s public — not whether it’s usable. Below: every signal we checked, what’s costing the score, and what to change.

Discry your API — freeView the docs ↗

SCORED UNDER RUBRIC 1.2 · A full re-launch under Discry Score 2.5 — a new behavioral instrument, not comparable to these scores — is in progress.

Discovery

45% of score · 60/100
OpenAPI specA machine-readable OpenAPI/Swagger spec agents can parse.Pass
llms.txtAn llms.txt index that points agents to the docs that matter.Fail
llms.txt qualityThe llms.txt is focused, current, and well under the size budget.Fail
llms-full.txtA full-text bundle agents can load in one request.Fail
AGENTS.mdAn AGENTS.md that tells coding agents how to build on the API.Partial
.well-known/mcp.jsonA discoverable MCP manifest at a well-known path.Fail
MCP registryThe API is listed in a public MCP registry.Pass
robots.txt AI directivesrobots.txt allows (or explicitly guides) AI crawlers.Pass
SitemapA sitemap so agents can enumerate the docs surface.Pass

Comprehension

55% of score · 91/100
Task-oriented descriptionsEndpoints described by what they accomplish, not just their shape.Pass
Realistic examplesRunnable, real-world request/response examples.Pass
Multi-step workflowsDocs that chain calls into complete jobs an agent can follow.Pass
Error-recovery guidanceDocumented failure modes and how to recover from them.Partial
Answer-first formatThe answer leads; preamble does not bury it.Pass
Capability boundariesClear limits — what the API can and cannot do.Pass
Naming consistencyConsistent, predictable naming across endpoints.Pass
Heading hierarchyClean heading structure agents can navigate.Pass
Markdown docsDocs available as clean markdown, not JS-rendered HTML only.Pass
Token efficiencyDocs are concise enough to fit an agent context window.Partial

What we found

  • An agent that reaches the docs gets near-ideal comprehension: every endpoint page leads with the HTTP method and path, uses task-oriented headings ('Create a Workspace', 'Lock a Workspace'), and ships copy-pasteable curl examples with realistic ws-/org IDs and JSON:API payloads (91/100 comprehension).
  • Discovery is the weak link: there is no llms.txt, no llms-full.txt, and no /.well-known/mcp.json, so an agent has no fast-path index and must crawl the 229 api-docs pages in the sitemap to learn what the API can do.
  • HashiCorp clearly invests in agent tooling — an official Terraform MCP server (listed on Glama, PulseMCP, and the official MCP registry) plus a dedicated hashicorp/agent-skills repo of Claude Code plugins — but this discoverability lives off-domain rather than in machine-readable files on developer.hashicorp.com.
  • robots.txt is agent-friendly (all crawlers allowed except GPTBot on one /validated-designs path), and the docs are MDX-sourced so they convert to clean markdown reliably.
  • The flagship HCP Terraform API is documented in prose (JSON:API), not as a downloadable OpenAPI spec; publicly accessible specs exist for sibling products (e.g. Boundary's swagger.json), so an agent's spec experience is inconsistent across the product family.

What to change

Prioritized by impact on discoverability. You (or your docs platform) deploy these — Discry never touches your API.

  1. 01Publish an llms.txt at developer.hashicorp.com pointing to each product's API-docs landing page (Terraform, Vault, Consul, Nomad, Boundary, HCP) with one-line descriptions — the single highest-leverage discovery fix, converting a crawl into a lookup.
  2. 02Add an llms-full.txt (or per-product markdown bundles) so agents can ingest core API concepts — auth, pagination, rate limits, JSON:API formatting — in one fetch instead of many page loads.
  3. 03Ship a downloadable OpenAPI 3.x spec for the HCP Terraform API (Boundary already publishes swagger.json) so agents can generate typed clients rather than parsing prose parameter tables.
  4. 04Add explicit recovery guidance to the status-code tables — e.g. 'on 429, honor x-ratelimit-limit / retry after backoff' and 'on 409 workspace-managing-resources, call the force-delete endpoint' — so agents can self-correct without a human.
  5. 05Add a canonical AGENTS.md to the primary product repos (terraform-mcp-server, terraform, vault) rather than only shipping an example-AGENTS.md template, giving coding agents in-repo context by convention.

Execution coverage · INFORMATIONAL, UNSCORED

Whether an agent can actually complete a call and recover from errors is the deeper Audit layer — documented here, but not part of the Discry Score.

Bearer API token (user / team / organization / audit-trails)OAuth (VCS integrations)✓ Error format documented✓ Rate limits documented✓ Pagination documented✕ Idempotency documented

Execution characteristics are documented thoroughly for the HCP Terraform API: bearer-token auth with four scoped token types, JSON:API error objects, explicit 30 req/s rate limits with per-endpoint overrides and an x-ratelimit-limit header, and page[number]/page[size] pagination. Idempotency keys are not documented.

See your own Discry Score.

Drop your API docs URL. See what an agent sees — in 60 seconds, free.

Discry your API — free