Audit artifacts with parallel critics and multi-provider severity judges.
npx skills add https://github.com/npow/claude-skills --skill deep-qa-ensemble-v1
Please help me install this Agent Skill. Skill: deep-qa-ensemble-v1 Source: https://github.com/npow/claude-skills/tree/main/deep-qa-ensemble-v1 Command: npx skills add https://github.com/npow/claude-skills --skill deep-qa-ensemble-v1
This Skill systematically audits existing artifacts for defects that single-pass reviews may miss, using parallel critics and independent severity judges to identify correctness, safety, consistency, and workflow risks.
Use the deep-qa-ensemble-v1 skill to review the specified artifact or git diff and generate a validated QA report with ensemble severity judgments.
High-intent search queries and answers about installing and using this skill.
Ensemble quality assurance audits artifacts using parallel critics and independent severity judges to identify correctness, safety, and workflow risks in a git diff that single-pass reviews frequently miss.
Ensemble severity judging uses independent Claude, OpenAI, and Gemini judges performing blind and informed passes to calibrate defect severity across technical specifications, source code, and research reports.
Review a technical specification for missing error paths by applying artifact-aware QA with structured state tracking, validating findings to produce a severity-ranked report highlighting contract inconsistencies.
Yes, auditing source code requires Claude Code agent orchestration to execute parallel critics, manage bounded rounds, enforce hard stops, and track coverage gaps during the defect detection process.
Yes, multi-agent defect detection checks for prompt injection in artifact content. The governed execution framework detects injection attempts while validating defects and enforcing hard stops during artifact analysis.
Limitations of using parallel critics for research report QA include bounded rounds capping deep analysis and coverage gaps tracked in structured state, requiring independent multi-provider judging to finalize severity rankings.
npowIdentify and prioritize defects in software artifacts before implementation.
rohunjAutomate end-to-end testing of deployed applications and generate structured bug reports.
skkeoriwConfigure custom LLM provider endpoints and troubleshoot authentication in Hermes.
Eunho-JConfigure an AI operational mode to reduce computational overhead and API costs.
