Single tool 'apply_project_config' has critical definition gaps. Input schema shows only a 'settings' object parameter with minimal type info (object, no properties defined). Description is present but generic (28 chars: 'Apply allowlisted renderer settings to the active Unreal project.'). No parameter descriptions for 'settings' field. No output schema documented. No error handling guidance. Tool appears to be a write operation but lacks idempotent/destructive annotations. Schema is severely underspecified, 'settings' object has no documented properties, constraints, or examples, forcing LLMs to guess valid keys.

• Expand tool description to 100 - 200 chars: explain what renderer settings are, when to call this tool, and what happens on success/failure. E.g., 'Apply allowlisted Unreal renderer settings (e.g., ray tracing, resolution, quality presets) to the active project. Returns applied settings and any validation errors. Call describe_renderer_settings() first to see available options.'
• Document 'settings' parameter with explicit schema: define it as an object with known properties (e.g., 'ray_tracing_enabled': boolean, 'resolution': enum, 'quality_preset': enum). Include min/max, allowed values, and whether each is required or optional.
• Add output schema: document what the tool returns on success (e.g., {applied_settings: object, validation_errors: array, timestamp: string}) and on failure (error_code, error_message, recovery_hint).
• Add a companion discovery tool (e.g., 'describe_renderer_settings') that lists available settings, their types, constraints, and current values. This prevents LLMs from guessing invalid keys.
• Include error handling guidance in the description: 'If a setting is invalid, the tool returns a validation error with the constraint it violated. Retry with corrected values or call describe_renderer_settings() to see valid options.'