Critical gaps across naming, descriptions, and schemas. Tool names lack consistent verb prefixes (snap_and_parent, tool_execute, product_shot violate single-responsibility). Descriptions are present but many are vague or in Spanish without English fallback. Input schemas visible for most tools but lack parameter descriptions, critical for LLM reasoning. No output schemas documented. Error handling absent. Security concerns: execute_blender_code and tool_execute are destructive but lack confirmation patterns or dry-run support.

• Split snap_and_parent into two tools: snap_objects(obj_move, obj_target, anchor_move, anchor_target) and parent_objects(child, parent). Each tool should have one clear responsibility.
• Replace tool_execute with explicit, named tools for the 239 registry items. If that's infeasible, document tool_execute's output schema and add a discovery tool (list_available_tools) so LLMs know what's callable.
• Add parameter descriptions to all tools. For snap_and_parent, document anchor format: 'One of: FRONT_BOTTOM_LEFT, FRONT_BOTTOM_RIGHT, ..., CENTROID (see addon/anchor_system.py:ANCHOR_NAMES)'. For product_shot/archviz/simple_scene, specify valid style/mood enums.
• Document output schemas for all tools in mcp_server.py or a separate schema file. E.g., get_scene_info should return {objects: [{name, type, id}], total_count}. This enables LLMs to chain tools correctly.
• Add dry-run or confirmation pattern to execute_blender_code. E.g., add a 'dry_run' boolean param that returns the code to be executed without running it, or require a separate confirm_execution tool. Cite code_guard validation in the description.

Inferred effective spec: <=2025-11-25.

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