16 tools with consistent naming (verb_noun pattern: check_, get_) and basic descriptions (avg ~80 chars). All tools have single-parameter schemas with type and description. However, descriptions are minimal and lack context on WHEN to use each tool vs. alternatives. No output schemas documented. No error handling guidance. Parameters lack constraints (e.g., company_name accepts any string; no validation hints). Tools are well-composed (one responsibility each) but descriptions don't explain dependencies or recovery paths.

• Document output schemas for all 16 tools. Specify field names, types, and whether results are paginated. Example: 'Returns {sponsorship_signal: string, confidence: number, data_points: integer}'.
• Expand descriptions to 100-150 chars explaining WHEN to use each tool. Example: 'Get salary statistics for a company. Use this for salary negotiation research; use get_company_full_profile for comprehensive company analysis.'
• Add validation hints to company_name, job_title, and location parameters. Example: 'Exact company name (e.g., "Google") or partial match. If ambiguous, tool returns top match or error with alternatives.'
• Add error recovery guidance to check_* tools. Example: 'If H1BGrader blocks lookup, wait 60 seconds and retry. If human verification required, error message will indicate next steps.'
• Document result limits and pagination. Example: 'Returns top 20 employers. Use offset/limit parameters (if supported) to paginate through results.'

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

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