Two tools with minimal but present descriptions and basic parameter schemas. Both tools have clear verb-noun naming (local_search, entity_lookup) and accept single string parameters. However, descriptions are terse (under 100 chars), lack context on when to use each tool, and do not explain output structure. No parameter descriptions beyond the query field itself. Output schemas are inferred from code (list of dicts with id/title/excerpt/score for search; name/kind/description/related/score for lookup) but not formally documented. Error handling is minimal, PermissionError raised but no guidance on recovery. No pagination support despite returning lists. Security is reasonable (read-only, allowlisted) but not explicitly documented in tool metadata.

• Add parameter descriptions: 'query (string): Natural language search query. Examples: "machine learning algorithms", "John Smith contact info". Max 500 chars.'
• Document output schemas in tool descriptions or via JSON Schema. E.g., 'Returns list of documents with fields: id (string), title (string), excerpt (string), score (float 0-1).'
• Expand tool descriptions to 100-200 chars explaining when to use each. E.g., local_search: 'Search document content by keyword or phrase. Use for broad topic searches.' entity_lookup: 'Find entities (people, places, concepts) and their relationships. Use when you need structured entity data.'
• Add limit and offset parameters to both tools with defaults (limit=20, offset=0) and document max result count.
• Enhance error messages: Instead of bare PermissionError, return structured error with recovery hint: '{"error": "local_search is disabled", "hint": "Contact admin to enable MCP_ALLOWED_TOOLS"}'