Six tools with basic GraphQL/REST wrappers around GitHub CLI. All tools have descriptions (10-200 chars, within baseline range) and input schemas with type definitions. However, critical gaps: (1) No output schemas documented, LLMs cannot predict return structure for chaining; (2) Parameter descriptions lack constraints (e.g., 'state' accepts OPEN|CLOSED|MERGED but no enum); (3) No error handling guidance, failures return raw JSON or stderr with no recovery hints; (4) Naming is verb-prefixed but overly specific (gh_graphql_* vs get_repo_info), adds noise without clarity; (5) No pagination support despite GraphQL queries returning lists; (6) Security: GraphQL queries built via string interpolation (owner_repo.split('/')), vulnerable to injection if input not validated. Average tool score: 52/100.

• Document output schemas for all tools. Example for gh_graphql_repo_info: {"name": string, "stargazerCount": integer, "forkCount": integer, "languages": [{"name": string}], ...}. This enables LLMs to plan downstream calls.
• Add enum constraints to 'state' parameters. Change description to formal constraint: 'state (enum: OPEN, CLOSED, MERGED; default: OPEN)' and update schema with 'enum' field.
• Implement input validation: validate owner_repo format before split(), reject special chars in usernames. Return actionable errors: 'Invalid repo format: expected "owner/repo", got "foo/bar/baz"'.
• Add error classification to run_gh_command(). Return {"success": false, "error": "...", "category": "retryable|user_fixable|fatal", "suggestion": "Try search_repos() first"}.
• Add pagination: accept 'limit' and 'cursor' params, return 'total_count' and 'next_cursor' in response. Cap default limit to 20 and document in description.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry