This server has 5 tools with basic schemas and descriptions, but significant quality gaps. Tool names follow verb_noun convention (generate_image, edit_image, get_time, web_search, web_scrape), which is good. However, descriptions are inconsistent in depth and parameter documentation is sparse. Most parameters lack type constraints (enums, ranges, patterns). No output schemas are documented. Error handling guidance is absent. The codebase shows partial implementation (image_tool.py snippet is truncated), making full assessment difficult. No evidence of idempotency, pagination, or recovery guidance.

• Document output schemas for all tools. Specify field names, types, and whether results are paginated. E.g., web_search should return {results: [{title, url, snippet}], total_count, next_cursor}.
• Add enum constraints to model_override and output parameters. Replace 'Optional ChatModel override' with 'One of: openai, azure, gemini' or link to a discovery tool.
• Add min/max bounds to numeric parameters. E.g., max_results: 'integer, 1 - 100, default 10' and latitude/longitude: 'number, -90 to 90'.
• Document mutual exclusivity and parameter precedence. E.g., 'Pass either query (single) or queries (array), not both. If both provided, queries takes precedence.'
• Add error recovery guidance to each tool description. E.g., 'If image generation fails due to content policy, the error will indicate the violation. Retry with a modified prompt or contact support.'

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

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