Three tools with basic schemas and descriptions, but significant gaps in parameter documentation and output structure. Tool names follow verb_noun convention (good), but descriptions lack actionable context for LLM selection. Parameter descriptions are minimal or absent. No output schemas documented. Error handling is absent. Schemas are present as JSON strings but lack depth in parameter guidance.

• Document output schema for each tool. Example for pixeldojo_generate_image: 'Returns {images: [{url: string, seed: int, model: string}], credits_used: int, total_credits_remaining: int}'. This enables chaining and LLM planning.
• Add parameter descriptions to all fields. For 'seed': 'Integer seed for reproducible image generation. Same seed + prompt + model produces identical output. Omit for random results.' For 'model_id': 'Optional model identifier. If omitted, returns all available models. Valid values: flux-pro, flux-1.1-pro, flux-1.1-pro-ultra, flux-dev-single-lora.'
• Enhance tool descriptions with discovery guidance. pixeldojo_describe_models: 'Call this first to see available models, their capabilities, and credit costs before generating images. Returns model details needed to choose the right model for your use case.'
• Add error handling examples in docstrings. E.g., 'If credits are insufficient, returns error with current balance and cost estimate. Retryable: no. Suggest checking credits with pixeldojo_get_credits first.'
• Implement structured error responses with recovery hints. Instead of raw exceptions, return JSON: {error: 'insufficient_credits', message: 'Need 50 credits, have 10', next_step: 'Call pixeldojo_get_credits to check balance or purchase more credits'}.

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

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