Score, lint, and refactor AI skill collections through iterative peer review.
npx skills add https://github.com/iuliandita/skills --skill skill-refiner-iuliandita
Please help me install this Agent Skill. Skill: skill-refiner Source: https://github.com/iuliandita/skills/tree/main/skills/skill-refiner Command: npx skills add https://github.com/iuliandita/skills --skill skill-refiner-iuliandita
This skill solves the problem of skill degradation and inconsistency in large collections by automating the scoring, linting, and improvement of individual skills through an adversarial, iterative loop.
Run the skill refiner to perform a quality sweep on the entire collection with a maximum of ten iterations.
High-intent search queries and answers about installing and using this skill.
Automated quality assurance for AI skill collections is achieved by running an iterative self-improvement loop that scores, lints, and refactors skills based on structural validation and behavioral testing metrics.
Cross-model peer review in automated testing uses secondary AI harnesses to adversarially verify skills, catching blind spots and preventing score inflation during the iterative refactoring loop.
Git is required to run iterative quality loops on skill directories, enabling structural validation, behavioral testing, and targeted refactoring changes across your AI skill collection.
Preventing skill degradation in large AI collections is best handled by targeting the lowest-scoring skills first and applying refactoring changes only when they measurably improve behavioral and structural metrics.
Skill-refiner requires integration with skill-creator to execute its automated scoring, linting, and adaptive improvement loop across targeted AI skill directories.
You should avoid using an iterative refactoring loop for AI agents when your project lacks secondary AI harnesses for cross-model peer review or when skills cannot be validated through automated behavioral testing.
nuoyanruoshuiValidate skill files for structural compliance and behavioral correctness.
bugrabilgeDesign benchmark test suites for evaluating LLM agent reliability.
awinogradovRun Google Gemini CLI code analysis, refactoring, and automated editing tasks.
muratcankoylanCreate deterministic self-verification checks for AI coding agents.
