Critical failures across all 10 tools. Every tool has empty input schemas (properties: {}) with zero parameter definitions. 8 of 10 tools lack any description text. The 2 tools with descriptions (cqs_index, cqs_wait_fresh) have minimal guidance. No output schemas documented. Tool names follow verb_noun convention (positive), but without schemas and descriptions, LLMs cannot determine parameter requirements or expected outputs. This server is non-functional for agent use.

• Add descriptions to all 9 undescribed tools (cqs_search, cqs_callers, cqs_notes_add, cqs_notes_remove, cqs_test_map, cqs_context, cqs_explain, cqs_wait_fresh, cqs_status). Each should be 50 - 200 characters, stating WHAT it does, WHEN to use it, and what it returns.
• Define input schemas for all tools. cqs_search needs parameters like 'query' (string, required), 'limit' (integer, 1 - 100, default 20). cqs_callers needs 'function_name' or 'symbol_id'. Document each parameter with type, description, constraints, and whether it is required.
• Document output schemas for all tools. Specify the structure of returned objects, field types, and which fields are safe to pass to downstream tools (e.g., if cqs_search returns 'file_id', ensure cqs_context accepts 'file_id').
• Add parameter descriptions to cqs_index. Clarify the relationship between 'overlay' and 'no_overlay' (mutually exclusive?), explain 'slot' naming conventions, and document 'overlay_root' validation rules.
• Add error handling guidance to all tools. Document common failure modes (e.g., 'index not ready', 'symbol not found'), classify them as retryable or user-fixable, and suggest recovery steps (e.g., 'Call cqs_wait_fresh() first').

Inferred effective spec: 2025-06-18+.

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