Three tools with Chinese descriptions and basic English translations. All have input schemas with types and descriptions, but descriptions are verbose (300+ chars) and mix implementation details with user guidance. Tool names are action verbs (search_, web_search_, read_) but descriptions lack clarity on when to use each vs. alternatives. No output schemas documented. Error handling is present but generic. Parameters have types and defaults but lack constraint documentation (e.g., limit max values not enforced in schema).

• Convert tool descriptions to 50 - 150 characters. Remove implementation details (token config, rate limits) and focus on user intent: 'search_github: Find GitHub repos by keyword (API, 60 req/hr without token)' → 'Search GitHub repositories by keyword, sorted by stars.'
• Document output schemas as JSON objects. search_github should return {repos: [{name, url, language, stars, description}], total_count, next_cursor} instead of formatted text.
• Add JSON Schema constraints: limit with maximum: 30, query with minLength: 1, maxLength: 256. Enforce in schema, not just description.
• Add tool selection guidance in descriptions. E.g., web_search_github: 'Use when search_github hits rate limits (free, no token needed).'
• Categorize errors: 'RateLimitError (retryable): GitHub API quota exhausted. Configure GITHUB_TOKEN or use web_search_github.' vs. 'NetworkError (retryable): Check connection.' vs. 'ParseError (fatal): GitHub page structure changed.'
• Split search_github description into two sentences: (1) what it does, (2) when to use it. Current description is 400+ chars and mixes concerns.

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

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