Two tools with minimal schema documentation and weak descriptions. analyzeImage has a basic schema with three parameters (image_base64, prompt, model) but descriptions are generic and lack actionable constraints. captureScreenshot has only two parameters with similarly generic descriptions. Neither tool documents output structure, error cases, or recovery guidance. No parameter validation rules, enums, or format constraints visible. Tool names are reasonable (verb_noun pattern) but descriptions fail to explain WHEN to use each tool or what distinguishes them. No evidence of idempotency, pagination, or composition patterns.

• Document complete output schemas for both tools. Example: analyzeImage returns {analysis: string, model_used: string, tokens_used: number}. captureScreenshot returns {screenshot_base64: string, analysis: string, dimensions: {width: number, height: number}}.
• Add enum constraint to model parameter: enum: ['claude-3-5-sonnet', 'claude-3-opus', 'gpt-4-vision']. Update description to list valid options.
• Expand parameter descriptions with format and validation rules. Example: 'image_base64: Base64-encoded image (JPEG/PNG, max 5MB, must be valid base64 string)'. 'prompt: User instruction for analysis (1-1000 chars, required for meaningful results)'.
• Add error handling guidance to tool descriptions. Example: 'Returns error if image is corrupted, too large, or model is unavailable. Retry with smaller image or different model if timeout occurs.'
• Distinguish tool purposes in descriptions. Example: analyzeImage: 'Analyze a provided image file using Claude Vision API. Use when you have an image URL or base64 data.' captureScreenshot: 'Capture current browser screenshot and analyze it. Use to inspect live web pages or desktop state. Requires browser-tools-mcp service running.'

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

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