validate-spec

Validate an OpenAPI Specification

Step 1: Run validation

Call the validate tool on the user's spec:

{"spec": {"file": "<path>"}}

If the user wants strict validation, add "strict": true.

⚠️ Tip for large specs: Use "no_warnings": true for initial triage. Large specs can produce hundreds of warnings that obscure the actual errors. Get the error picture first, then run again without no_warnings to review warnings separately.

Step 2: Report results

If the spec is valid:

• Confirm it passes validation
• Mention the detected OAS version
• Note any warnings (unless no_warnings was set)

If the spec has errors:

• List each error with its JSON path and a plain-language explanation
• Group related errors (e.g., multiple missing $ref targets)
• Explain why each error matters and what it would cause in practice (tooling failures, code generation issues, etc.)

Paginating large results

Results are paginated (default limit: 100). When returned < error_count, there are more errors:

{"spec": {"file": "<path>"}, "offset": 100, "limit": 100}

⚠️ Strategy for large error sets: Analyze the first page for patterns — if errors are repetitive (e.g., hundreds of missing path parameters), summarize the pattern and total count rather than paging through all of them. Only page further when errors appear diverse or the user needs specifics.

Step 3: Suggest fixes

For each error, explain how to fix it. Common patterns:

Error type	Suggested fix
Missing path parameter	Add the parameter to the operation's parameters array
Duplicate operationId	Rename to be unique, following a verb+resource pattern
Invalid $ref target	Fix the reference path, or stub the missing schema
Missing required field	Add the field with an appropriate value

If the errors are auto-fixable, offer to run the fix tool:

{"spec": {"file": "<path>"}, "dry_run": true}

Show the user the planned fixes before applying. ✅ When applying, use output to persist: