AegisDB MCP server has 5 tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names are action-oriented (recall, capture, ask, edit, delete) but descriptions are brief (10-50 chars, below the 50-200 char baseline for LLM optimization). Parameters lack descriptions for most fields, e.g., 'tags' array has no guidance on format or constraints. Output schemas are not documented anywhere in the provided code. Error handling and recovery guidance are absent. The server is STDIO-only, which is a hard transport limitation.

• Expand tool descriptions to 50-200 chars. Example: 'recall' → 'Search stored memories by semantic similarity (embedding-based) or keyword match. Returns up to [limit] results with relevance scores. Use semantic=true for conceptual queries, false for exact phrase matching.'
• Add parameter descriptions for all fields. For 'tags': 'Array of string tags (max 5 per memory, 20 chars each). Used to filter results in recall() and ask(). Example: ["project-x", "urgent"]'
• Document output schemas for each tool. Example for recall(): 'Returns {results: [{id, text, kind, tags, importance, confidence, relevance_score}], total_count, next_cursor}'
• Add numeric constraints to descriptions. Example for 'limit': 'Maximum results to return (1-100, default 20). Larger limits may slow semantic search.'
• Implement error recovery guidance. Example: 'If semantic search fails, try keyword search (semantic=false). If memory ID not found, call recall() to discover valid IDs.'