code-review

Review uncommitted code changes and generate evidence-backed findings for approval.

6|16|Updated Mar 9, 2026
One-click install
npx skills add https://github.com/flightctl/ai-workflows --skill code-review-flightctl
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-review
Source: https://github.com/flightctl/ai-workflows/tree/main/code-review
Command: npx skills add https://github.com/flightctl/ai-workflows --skill code-review-flightctl

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you perform a structured, evidence-based AI code review of local uncommitted changes, so you can decide what to fix before committing or opening a PR.

Core Features & Use Cases

  • Interactive /start, /continue, /clean flow: Run an initial review, iteratively re-review after accepted fixes, and clean up artifacts when abandoned.
  • Reviewer profile-driven standards: Discovers project conventions (AGENTS.md/CLAUDE.md, contributing docs, lint/test/CI signals) and applies them to the review.
  • Human decision table + unattended mode: Presents findings with an implementor’s independent assessment and supports an automated iteration loop with safety guardrails.

Quick Start

Run the code-review workflow’s start command and decide on the presented findings table before implementing any accepted changes.

Frequently Asked Questions about code-review

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

FAQPage Schema
How do I run an AI code review on uncommitted changes before committing?

Run an interactive code review on uncommitted changes by starting the workflow to discover project conventions, generate evidence-backed findings with file/line references, and iteratively re-review after accepted fixes.

Can I automate pre-commit validation for local code changes?

Yes, you can automate pre-commit validation using an unattended mode with safety guardrails, which loops through accepted fixes automatically while enforcing user-controlled decisions for escalations.

How does AI code review discover and apply my project conventions?

AI code review discovers project conventions by reading AGENTS.md or CLAUDE.md files, contributing docs, and lint or CI signals, then applies these standards to evaluate the uncommitted diff context.

What's the best way to handle iterative feedback after fixing code review findings?

Use the interactive continue flow to handle iterative feedback, which re-reviews uncommitted changes after you implement accepted fixes, ensuring all actionable findings are resolved before committing.

Does this code review tool work without external dependencies?

Yes, this code review tool works without external dependencies, operating entirely as a structured workflow that reads local diff context and writes review artifacts for human approval.

When should I clean up artifacts from a local pre-PR code review?

Clean up review artifacts when abandoning a local pre-PR code review by running the clean command, ensuring generated findings and workflow files do not clutter the project directory.