a2db has well-structured tool definitions with clear naming (verb-first: login, logout, list_connections, execute, search_objects) and comprehensive parameter schemas. All 5 tools have descriptions and typed input schemas. However, output schemas are not documented, LLMs cannot predict response structure for downstream chaining. Parameter descriptions are present but generic in places (e.g., 'Connection spec' lacks detail on required fields). Error handling guidance is absent, tools do not indicate what to do on failure. Security is reasonable (no secrets in params) but audit logging is not mentioned.

• Document output schemas for all tools. Example for list_connections: 'Returns array of {project: string, env: string, db: string, db_type: string}. No secrets included.'
• Add error recovery guidance to tool descriptions. Example: 'If login fails with "Connection refused", verify DSN format and network access. Call list_connections to see valid saved connections.'
• Clarify the 'connection' parameter: is it required? What is the default if omitted? Does it override a session default? Add example: {project: 'myapp', env: 'prod', db: 'users'}.
• Document execute 'queries' parameter structure: 'Pass object {query_name: sql_string}, array [sql1, sql2] (auto-named q1, q2), or JSON string. Example: {users: "SELECT * FROM users LIMIT 10"}'.
• Expand search_objects 'detail_level' description: 'names: column/table names only. summary: names + row count, column types. full: complete DDL, indexes, constraints. Use names for discovery, full for schema export.'