review-loop

Coordinate automated review-fix cycles for Codex app-server threads via CLI.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/bytes032/agentkit --skill review-loop-bytes032
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-loop
Source: https://github.com/bytes032/agentkit/tree/main/skills/review-loop
Command: npx skills add https://github.com/bytes032/agentkit --skill review-loop-bytes032

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Codex app-server review loops often require manual triage, repetitive fixes, and repeated validations. This Skill automates end-to-end review→fix cycles, accelerating iteration and improving code quality across threads.

Core Features & Use Cases

  • Automated review loops: orchestrates deterministic cycles using a thread id to identify and apply fixes.
  • Flexible targets: supports base, commit, uncommitted, and custom instructions for focused reviews.
  • Output options: provides human-readable, JSON Lines, or plain text outputs to fit automation pipelines and dashboards.

Quick Start

Run codex-review-loop with a thread identifier to start the review→fix cycle against a Codex app-server thread.

Frequently Asked Questions about review-loop

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

FAQPage Schema
How do I automate Codex review loops for app-server threads?

Automating Codex review loops involves using a CLI workflow with an explicit thread id to orchestrate end-to-end review and fix cycles. This Skill coordinates deterministic cycles against base, commit, or uncommitted targets to accelerate iteration.

Can I output Codex code review results as JSON for automation pipelines?

Yes, Codex code review results support JSON Lines and plain text outputs alongside human-readable formats. These output options allow you to route review loop data directly into automation pipelines and dashboards.

How do I run a code review cycle against uncommitted changes in Codex?

Running a code review cycle against uncommitted changes requires specifying the uncommitted target mode via the CLI. The workflow uses a thread identifier to apply fixes and validations specifically to your uncommitted code state.

Does the Codex review loop CLI require a specific thread id to start?

Yes, the Codex review loop CLI requires an explicit thread id to identify the correct app-server thread. This identifier ensures the automated review and fix cycle applies to the correct thread sourced from the Codex UI or prior CLI output.

What are the limitations of automating code review loops with Codex?

Automating code review loops with Codex operates within safety defaults to prevent unintended modifications. While it supports custom instructions and multiple target modes, complex triage scenarios may still require manual validation outside the automated cycle.