◎ Discry Score
gitlab.com
devtools · API
C
0 / 100
DISCOVERY0
COMPREHENSION0
Category leader: 96 (A)
Discry your API →
DEVTOOLS · RANK #22 OF 39

GitLab is Needs Work to agents.

Discry independently scored how well an AI agent can discover and understand the GitLab 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 · 74/100
OpenAPI specA machine-readable OpenAPI/Swagger spec agents can parse.Partial
llms.txtAn llms.txt index that points agents to the docs that matter.Pass
llms.txt qualityThe llms.txt is focused, current, and well under the size budget.Partial
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.Pass
.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 · 71/100
Task-oriented descriptionsEndpoints described by what they accomplish, not just their shape.Partial
Realistic examplesRunnable, real-world request/response examples.Partial
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.Partial
Token efficiencyDocs are concise enough to fit an agent context window.Partial

What we found

  • GitLab has a well-structured llms.txt (70KB, 42 sections) but it functions as a site-wide navigation dump rather than an API-focused guide — an agent looking for API patterns would need to sift through product documentation sections to find API content
  • The OpenAPI spec exists but only covers a small subset of endpoints — GitLab's own issue tracker acknowledges the spec is incomplete, which limits agent tooling that relies on OpenAPI for auto-discovery
  • AGENTS.md exists in the gitlab-runner repo with useful coding agent guidance, and GitLab is well-represented in MCP registries with an official Anthropic-maintained MCP server on PulseMCP
  • Error documentation provides clear status code tables with descriptions but recovery steps are generic — an agent encountering a 422 would know 'entity couldn't be processed' but not which specific field or validation failed
  • Comprehensive tutorial section with multi-step workflow guides and the REST API overview page provides excellent structured context for pagination, authentication, and request formatting

What to change

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

  1. 01Complete the OpenAPI specification to cover all REST API endpoints — this is the single highest-impact improvement for agent discovery (weight: 5)
  2. 02Create a focused llms.txt under 50KB that covers API capabilities specifically, linking to REST API, GraphQL, webhooks, and authentication — separate from the product navigation dump
  3. 03Add llms-full.txt with comprehensive API documentation in markdown format (currently returns 403)
  4. 04Enhance error documentation with actionable recovery steps — for each 4xx status code, explain what specifically went wrong and how to fix it
  5. 05Add .well-known/mcp.json pointing to the official Anthropic-maintained GitLab MCP server

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.

personal_access_tokenoauth2project_access_tokengroup_access_token✓ Error format documented✓ Rate limits documented✓ Pagination documented✕ Idempotency documented

Comprehensive REST and GraphQL APIs with multiple auth methods. JSON responses with documented status codes. Rate limits documented with specific headers (RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset). Keyset and offset-based pagination documented. No idempotency keys.

See your own Discry Score.

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

Discry your API — free