jobmatch-mcp has 4 tools with clear verb-noun naming (search_jobs, get_job_details, rank_jobs_for_candidate, list_sources). All tools have descriptions (10 - 200 chars, within baseline). Input schemas are present with type declarations and parameter descriptions. However, output schemas are NOT documented, the code shows tool registration but no explicit return type documentation visible in source. The rank_jobs_for_candidate tool has a TODO comment indicating incomplete implementation. Error handling is minimal; no recovery guidance or actionable error messages. Parameter constraints (e.g., limit max:50) are declared but not validated in visible code. Overall: solid naming and basic schema structure, but missing output documentation and error handling patterns.

• Document output schemas for all tools. Specify return types, field names, and structure. Example: search_jobs returns array of {id, title, company, location, url, score?}. This enables LLM to chain tools and extract required fields.
• Implement rank_jobs_for_candidate ranking logic. Currently returns score=0 for all jobs. Implement profile-aware scoring using skills alignment, title match, location preference, and seniority level as noted in TODO.
• Add error handling with recovery guidance. Wrap HTTP calls in try-catch. Return structured errors: {error: 'string', code: 'enum', retryable: boolean, suggestion: 'string'}. Example: 'API key missing. Configure JOBMATCH_THEIRSTACK_API_KEY environment variable.'
• Validate input parameters before processing. Check limit ≤ 50, query is non-empty, sourceIds are known. Return clear validation errors: 'Invalid limit: 100 (max 50)' instead of silent failures.
• Enhance list_sources description. Explain: 'Returns available job sources and their status (active/disabled). Call this first to discover which sources are configured and ready to search.'