skill-quality

Classify skill contract failures from eval results and edit SKILL.md definitions.

Updated Mar 1, 2026
One-click install
npx skills add https://github.com/adhi-jp/agent-skills --skill skill-quality-adhi-jp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-quality
Source: https://github.com/adhi-jp/agent-skills/tree/main/skills/skill-quality
Command: npx skills add https://github.com/adhi-jp/agent-skills --skill skill-quality-adhi-jp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Translate failures and feedback into contract changes that improve skill and eval reliability, ensuring decisions are grounded in evidence.

Core Features & Use Cases

  • Evidence intake and classification to guide targeted SKILL.md edits, references, and eval adjustments.
  • Small, repeatable contract deltas that close gaps without overhauling entire workflows.
  • Structured evaluation design and outcome analysis to validate improvements and guard against regressions.

Quick Start

Identify the observed failure, map it to a reusable contract dimension, and implement the smallest change that closes the gap.

Frequently Asked Questions about skill-quality

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix skill contract failures using evidence from eval results?

Map the observed failure to a reusable contract dimension and implement the smallest change that closes the gap. This structured approach edits SKILL.md or references without overhauling entire workflows.

What is a contract delta in skill quality assurance?

A contract delta is a small, repeatable modification applied to SKILL.md or eval definitions to close reliability gaps. It targets specific failure dimensions identified from evidence signals.

How do I design evals to prevent skill regressions across configurations?

Update eval definitions in the same change set as contract edits, then run the shared eval workflow to validate improvements and guard against regressions across with_skill and without_skill configurations.

Do I need reviewer notes and session history to classify skill failures?

No, but combining reviewer notes and session history with eval results provides a stronger evidence-based foundation for classifying failures and guiding targeted SKILL.md edits.

What's the best way to update eval definitions when modifying a skill contract?

Update eval definitions in the same change set as the contract edits. This ensures the shared eval workflow validates improvements across both with_skill and without_skill configurations simultaneously.