critic

Validate AI output against a contract with a 0.70 confidence threshold.

1|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/jimmymalhan/codereview-pilot --skill critic-jimmymalhan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: critic
Source: https://github.com/jimmymalhan/codereview-pilot/tree/main/.claude/skills/critic
Command: npx skills add https://github.com/jimmymalhan/codereview-pilot --skill critic-jimmymalhan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill ensures that AI-generated output meets strict quality standards, preventing the use of incomplete or low-confidence results in critical workflows.

Core Features & Use Cases

  • Output Validation: Verifies that all required fields (root cause, evidence, fix plan, rollback, tests, confidence) are present and valid.
  • Confidence Threshold: Blocks output if the confidence score is below 0.70.
  • Use Case: Use this Skill as a final check before deploying a fix suggested by an AI, ensuring it's reliable and actionable.

Quick Start

Use the critic skill to validate the AI's output for the incident report.

Frequently Asked Questions about critic

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

FAQPage Schema
How do I validate AI-generated output before deploying a fix?

To validate AI-generated output, use a quality gate to verify the response includes a root cause, evidence, fix plan, rollback, and tests. This ensures the output is complete and reliable before deployment.

What is a confidence check for AI output and how does it work?

A confidence check for AI output enforces a minimum score threshold, such as 0.70, to block low-certainty results. This prevents incomplete or unverified AI workflows from being used in critical deployments.

How do I enforce a contract on AI incident reports?

You enforce a contract on AI incident reports by validating the output against predefined required fields. This checks for the presence of root cause, evidence, fix plan, rollback, tests, and confidence to ensure quality.

Can I block AI workflows if the confidence score is below 0.70?

Yes, you can block AI workflows if the confidence score is below 0.70 by applying a quality gate. This stops incomplete or low-confidence results from being used in critical workflows requiring evidence-based outputs.

What are the limitations of using a quality gate for AI output validation?

A limitation of a quality gate for AI output validation is that it strictly enforces required fields and a 0.70 confidence threshold. Avoid this approach if your AI workflow does not require evidence-based results or high confidence scores.