Server has 9 tools with basic schemas and descriptions, but significant quality gaps. Descriptions are present but often generic or in Chinese without English translations for LLM clarity. Parameter descriptions are minimal. No output schemas documented. Error handling is absent, tools fail silently or return raw exceptions. Tool names lack consistent verb-noun patterns. No pagination support despite database queries. Security concerns: credentials in environment variables (good), but no input validation against SQL injection despite accepting raw SQL queries. Composition is weak, tools are loosely related (database + Ivanti API + RAG) without clear chaining support.

• Translate all descriptions to English and expand to 50 - 200 characters. Include WHAT the tool does, WHEN to use it, and what it returns. E.g., 'query_mysql: Execute a SELECT query against the ITSM database. Use this to fetch records by custom criteria. Returns an array of row objects with column names as keys.'
• Document output schemas for every tool. Specify field names, types, and structure. E.g., list_tables returns [{table_name: string, row_count: int, size_mb: float}].
• Replace f-string SQL construction with parameterized queries. Use aiomysql's cursor.execute(query, params) pattern throughout. Validate table/column names against a whitelist before interpolation.
• Add error handling with actionable messages. E.g., 'login_ivanti failed: Invalid credentials. Verify tenant, username, password, and role.' or 'query_mysql timed out after 30s. Try a simpler query or add a LIMIT clause.'
• Add pagination to 'list_tables' and 'query_mysql': accept limit (default 20, max 100) and offset parameters. Return total_count and has_more fields so LLMs can iterate.