Server has 4 tools with basic schemas and descriptions, but significant gaps in parameter documentation, output schema clarity, and error handling guidance. Tool names follow verb_noun convention (search_*, get_*, debug_*), which is good. However, descriptions are generic (avg ~80 chars, below 194-char baseline), parameters lack type constraints and validation rules, and output schemas are not documented. The 'filters' parameter in search_companies is an untyped object with no guidance on valid keys. search_by_funding has no required parameters despite needing at least min or max funding. Error handling returns raw API errors without recovery guidance. No tool annotations (readOnlyHint, etc.) despite all being read-only. This is typical community-grade work, functional but not production-ready.

• Add output schema documentation to each tool. E.g., search_companies returns [{company_id, name, domain, sector, funding_usd, stage, founded_year}]. This enables LLMs to chain calls and extract fields.
• Replace 'filters' object with explicit enum parameters: sector (enum: Cybersecurity, AI, FinTech, ...), stage (enum: Seed, Series A, Series B, ...), country (string). Untyped objects invite hallucinated keys.
• Make search_by_funding require at least one of min_funding or max_funding. Add validation: 'min_funding and max_funding must be positive integers in USD. If only min is set, returns companies funded >= min. If only max is set, returns companies funded <= max.'
• Enhance error messages with recovery steps. E.g., 'TRACXN_ACCESS_TOKEN not set. Set it via: export TRACXN_ACCESS_TOKEN=your_key' or 'Rate limited (429). Retry after 60 seconds.' or 'Domain not found. Try search_companies() to find the correct domain.'
• Add tool annotations to all tools: readOnlyHint=true for all four (they only read data, never modify). This signals to agents that these calls are safe to retry and have no side effects.

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

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