quality-check-scoped

Evaluate PR diffs against taxonomy rules and emit findings.yaml.

3|Updated Jan 25, 2026
One-click install
npx skills add https://github.com/kapilvirenahuja/garura --skill quality-check-scoped
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: quality-check-scoped
Source: https://github.com/kapilvirenahuja/garura/tree/main/core/components/skills/quality-check-scoped
Command: npx skills add https://github.com/kapilvirenahuja/garura --skill quality-check-scoped

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

quality-check-scoped eliminates slow, noisy PR review audits by evaluating only the quality standards that are relevant to the files and diff hunks actually touched by a pull request.

Core Features & Use Cases

  • Diff-bounded evaluation: Restricts reads and standard matching to changed_paths, ensuring the scan scope never expands beyond what the PR modified.
  • Taxonomy-driven determinism: Loads the PR severity taxonomy from severity_taxonomy_path and applies mechanical match rules to produce findings with taxonomy-sourced severities.
  • KB reuse without subagent execution: Reuses the same quality-check knowledge base data model, but performs the work as a single-pass evaluator and emits findings.yaml.
  • Production-ready output for tooling: Produces a stable, review-friendly findings.yaml with counts by severity and a computed scan_coverage metric.

Quick Start

Run the review-pr flow with the payload containing diff_path, changed_paths, standards_set, severity_taxonomy_path, intent_summary, and output_path to generate findings.yaml for the PR.

Frequently Asked Questions about quality-check-scoped

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

FAQPage Schema
What is diff-bounded quality evaluation for PR review?

Diff-bounded quality evaluation restricts PR review scans to only changed_paths and unified diff hunks, matching taxonomy rules to emit structured findings without scanning the entire repository.

How do I run a deterministic PR review on only changed files?

Provide a payload containing diff_path, changed_paths, standards_set, severity_taxonomy_path, intent_summary, and output_path to execute a single-pass evaluation that generates a structured findings.yaml file.

Does quality-check-scoped support full-repo scanning for quality standards?

No, it prohibits reads outside changed_paths to guarantee scan scope containment, performing diff-bounded evaluation strictly for enterprise PR review workflows requiring reproducible classification without full-repo scanning.

How does taxonomy matching ensure correct severity in PR review findings?

Taxonomy-driven determinism loads the PR severity taxonomy from severity_taxonomy_path and applies mechanical match rules to produce findings with taxonomy-sourced severities and correct standard_id classification.

What format does the PR review output use for downstream tooling integration?

It produces a stable, review-friendly findings.yaml containing counts by severity and a computed scan_coverage metric, enabling deterministic integration with enterprise PR review workflow tooling.

Why does my PR review fail pre-flight validation when running diff analysis?

Pre-flight validation requires a valid severity taxonomy, non-empty diff and changed_paths, and resolvable standards. If any are missing or invalid, the deterministic diff analysis halts before evaluation begins.