agent-auditor

Audit and fix QA pipeline agents by analyzing contracts and cross-file dependencies.

Updated Aug 23, 2026

Search Tags:#automation #qa #audit #agents #references #edits

Ready to run

npx skills add https://github.com/JuantoC/BlueStack-Test-Automation-Framework --skill agent-auditor-juantoc

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-auditor
Source: https://github.com/JuantoC/BlueStack-Test-Automation-Framework/tree/main/.claude/skills/agent-auditor
Command: npx skills add https://github.com/JuantoC/BlueStack-Test-Automation-Framework --skill agent-auditor-juantoc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Audita y corrige los cinco agentes del pipeline QA ubicados en .claude/agents/, detectando defectos, inconsistencias entre contratos, contenido externalizable y referencias rotas. Ejecuta fixes directos y escala al humano únicamente ante decisiones funcionales o arquitectónicas ambiguas.

Core Features & Use Cases

Audita en paralelo los cinco subagentes del pipeline QA (qa-orchestrator, ticket-analyst, test-engine, test-reporter, test-generator) y consolida hallazgos con evidencia.
Propone ediciones precisas para correcciones (mover contenido a references/, arreglar rutas rotas, normalizar contratos) y ejecuta los fixes cuando sea seguro.
Soporta escalación cuando se requieren decisiones humanas, describe las opciones y justificaciones para revisión.
Puede referenciar la detección-framework y wiki para guiar cambios consistentes.

Quick Start

Ejecuta una auditoría de los cinco subagentes en paralelo y revisa los hallazgos consolidados antes de cualquier ejecución.

Frequently Asked Questions about agent-auditor

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

FAQPage Schema

How do I audit and fix broken references across QA pipeline agents?▼

You audit QA pipeline agents by analyzing contracts, references, and cross-file dependencies in .claude/agents to detect defects and broken references. The process proposes precise edits, executes safe fixes directly, and escalates ambiguous architectural decisions to humans for review.

What is the best way to normalize contracts and externalize content for subagents?▼

Normalizing subagent contracts involves proposing precise edits to move externalizable content to references directories and fix broken paths. This ensures consistency across the five QA pipeline agents by aligning their contracts with the detection-framework and wiki guidelines.

Can I run parallel audits on multiple QA agents and consolidate the findings?▼

Yes, you can audit the five QA subagents in parallel and consolidate findings with evidence. This parallel approach detects inconsistencies across agents like qa-orchestrator, test-engine, and test-generator, providing a unified report before executing any corrections.

When should I escalate QA agent corrections instead of executing automatic fixes?▼

Escalation is required for functional or architectural decisions that are ambiguous and cannot be safely automated. The Skill describes available options and justifications for human review, executing direct fixes only when they are safe and fully deterministic.

Does the agent auditor require dependencies to detect cross-file misalignments?▼

No dependencies are required. The auditor operates directly on .claude/agents and associated references in the repository, reading the detection-framework and wiki to guide consistent changes and identify cross-file misalignments without external packages.