iterate-until-good

What problem does it solve?

This Skill provides a structured, repeatable review-and-improve workflow that evaluates code, pull requests, branches, or documents across multiple quality categories and drives iterative changes until each category exceeds an 85% threshold while treating scores below 70% as a hard block.

Core Features & Use Cases

• Multi-axis evaluation: Produce per-category 0-100 scores for simplicity, flexibility, usability, documentation and discoverability, security, pattern conformity, best practices, and general QA.
• Agentic iteration: Launch background review agents to run parallel evaluations, generate detailed reports, leave inline PR comments, and iterate on the target until quality thresholds are met.
• Auditable reports: Persist structured reports and supporting references into a consistent repository path for traceability and follow-up remediation.
• Use Case: Use this Skill on a feature branch or pull request to identify weak areas (e.g., documentation, security) and automatically drive iterative fixes until the branch meets merging quality standards.

Quick Start

Run the iterate-until-good skill on the target pull request or branch and iterate until every category is above 85%.