Dev Feedback Loop

关于

The Dev Feedback Loop skill is a rigorous continuous improvement system for AI-driven software engineering. It monitors development cycle effectiveness by calculating precise assertiveness scores based on iteration counts, review outcomes, and validation success. By performing root cause analysis on failures and systematically auditing agent prompt quality, it identifies recurring patterns and bottlenecks. This ensures that every task—whether a simple fix or a complex experiment—contributes data to a refined, high-quality development workflow within the docs/feedbacks directory.

主要功能

• Automated assertiveness score calculation using a weighted penalty matrix
• Agent prompt quality auditing with automated improvement suggestions
• Centralized performance reporting aggregated by date and specific AI agent
• 17 GitHub stars
• Systematic root cause analysis for development failures and bottlenecks
• Mandatory workflow tracking via TodoWrite to ensure metrics are never skipped

使用场景

• Enforcing high-quality software engineering gates through data-driven performance metrics
• Identifying systemic prompt failures to optimize agent instructions and rules
• Measuring and improving the efficiency of AI agents across multiple development cycles