Two tools with minimal documentation. interactive_feedback has a bilingual description (Chinese/English) but lacks parameter descriptions, the schema shows three input fields (interactive_feedback, command_logs, images) with only type and description keys, but no guidance on format, constraints, or when to use each. get_system_info has no parameters but its description is vague ('Get system environment information'). Neither tool documents output schema, error handling, or recovery paths. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) present. Naming is acceptable (verb_noun pattern), but descriptions fall short of the 50-200 char LLM-optimized baseline and lack actionable context.

• Add detailed parameter descriptions to interactive_feedback. For each of interactive_feedback, command_logs, and images, specify: expected format (plain text, JSON, file path?), constraints (max length, required vs optional), and when to include each. Example: 'interactive_feedback (string, required): User's natural-language feedback or complaint. Max 2000 chars.'
• Document output schema for both tools. Specify the response structure, field types, and what data is returned. Example for get_system_info: 'Returns {os: string, python_version: string, installed_packages: array<string>}'.
• Expand tool descriptions to 100-150 chars with explicit WHAT/WHEN/WHY. Example: 'Collect user feedback on command execution. Call this after running a command to capture user reactions, error logs, and screenshots. Returns feedback_id for tracking.'
• Add tool annotations. Mark interactive_feedback with destructiveHint=true (it writes feedback) and get_system_info with readOnlyHint=true. This enables agents to reason about retry safety and side effects.
• Add error handling guidance. Document common failure modes (e.g., 'If images array is too large, return error with max_size_mb and suggest splitting into multiple calls'). Include recovery hints so LLMs know whether to retry, ask the user, or abort.

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

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