Three search tools with consistent naming (verb_noun pattern) and basic descriptions. All tools have input schemas with type definitions and parameter descriptions. However, descriptions are minimal (50-80 chars, below the 194-char baseline), lack WHEN/WHY guidance, and output schemas are not documented. No error handling guidance, no pagination despite search tools returning lists, and no indication of result limits. Tools are read-only (good security posture) but lack the depth expected for production-grade agent integration.

• Expand tool descriptions to 150-250 chars. Include: what the tool searches (papers vs sections vs chunks), when to use it (broad discovery vs detailed evidence), and what structure is returned. Example: 'Search section-level summaries in the CivilMCP v2 index. Use this for structured evidence retrieval when you need specific findings from papers. Returns section_id, paper_title, summary, relevance_score.'
• Document output schemas explicitly. For each tool, specify: field names, types (string, number, array), and what each field represents. Include example response shape in docstring or separate schema file.
• Add pagination guidance. Specify: default limit (e.g., 20), max limit (e.g., 100), and whether cursor-based or offset-based pagination is supported. Update description: 'Returns up to 20 results by default; pass limit (1-100) to adjust.'
• Add error handling hints to descriptions. Example: 'If query is empty or too long, returns 400 with guidance. If database is unavailable, returns 503 with retry-after header.'
• Constrain the 'limit' parameter with min/max bounds in schema. Add description: 'Maximum results to return (1-100, default 20). Larger limits may timeout.'
• Consider splitting search_civil_knowledge into two tools if the use cases differ significantly: search_papers (broad discovery) and search_evidence (detailed retrieval). This reduces LLM confusion about which tool to call.

Inferred effective spec: 2026-07-28+.

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