13 tools with consistent naming (gd_* prefix) and action verbs, but critical gaps in descriptions and schemas. All tools have descriptions (10-50 chars), but most are terse and lack context on WHEN to use them or what they return. Input schemas present but incomplete: parameters lack type definitions for several fields (e.g., 'groups' is string but no format/pattern guidance). No output schemas documented. No error handling guidance. No parameter validation rules (e.g., what happens if x/y are negative? What's the valid range for objectId?). Composition is reasonable (each tool does one thing), but no idempotency guarantees or recovery guidance.

• Expand each tool description to 100-200 chars. Include: what it does, when to use it (vs similar tools), what it returns, and any prerequisites. E.g., 'Create a text object at (x, y) with optional scaling and group assignment. Returns object_id for use in triggers. Use before gd_move_trigger to animate text.'
• Document output schemas for all tools. Specify: success boolean, object_id (if created), position, group assignments, and any error details. This enables tool chaining.
• Add parameter constraints to descriptions: 'objectId: integer 1-999 (1=block, 2=spike, 40=platform, etc.)', 'x/y: float, any value', 'groups: colon-separated integers or "0" for none', 'opacity: float 0.0-1.0'.
• Add error handling guidance: 'If objectId is invalid, returns error with valid range. If group doesn't exist, returns warning but creates object anyway. Retry-safe: same params = same result.'
• Create a discovery tool (gd_list_object_types) that returns valid objectIds and their names. This lets agents self-correct invalid IDs without trial-and-error.