←Leaderboard/Code Quality/claudex-loop
Last commit on September 6, 2026·Created on June 5, 2026

chaseai-yt/claudex-loop

“A multi-model adversarial loop for hardening high-stakes architectural plans”
Combined rank
#162
across all skills
In Code Quality
#8
category rank
Stars
2.8k
+4.0% in last 7d
Forks
302
+8.2% in last 7d
Watchers
21
0.0% in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
TL;DR

Claudex Loop facilitates a two-act hardening process where one AI model drafts an implementation plan and another independently reviews it in a read-only sandbox. This adversarial approach is designed for technical leads and architects who need to stress-test critical system designs, such as authentication flows or database schemas, before writing code.

By coordinating a bounded revision loop between Claude and Codex, the system ensures that plans are rigorously cross-inspected and locked in a dedicated file to prevent implementation errors in complex environments.

WHO IT'S FOR
Technical leads / architects
stress-test high-stakes architectural plans before implementation
Backend / infrastructure engineers
validate complex system designs via adversarial AI review
Platform / DevEx teams
align technical plans with existing domain models
AI agent power users
implement multi-model verification workflows for coding
Repository contents

6 skill files

Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.

Use this skill collection

Follow the documented setup, then try a first task.

I’m using

Claude Code plugin bundle

Installs: Complete claudex-loop plugin bundle, including Claudex Route, Claudex Loop, codex-review, and codex-build. · Claude Code

Before you start
  • Claude Code with plugin marketplace support.
  • For the full cross-provider workflow: both the Codex and Claude CLIs must be installed and authenticated, and Python 3.10+ must be installed. No runtime pip packages or separate API keys are required.
  1. Enter these commands in a Claude Code agent session, not a terminal.

    /plugin marketplace add chaseai-yt/claudex-loop
    /plugin install claudex-loop@claudex-loop
README.md ↗ · Checked Oct 7, 2026.claude-plugin/marketplace.json ↗ · Checked Oct 7, 2026

Give it something to do.

Author’s example

Choose who should review a plan

Uses claudex-route

  • A plan or sufficient plan context in the current conversation
  • Any cost, speed, model, or provider constraints
claudex-route: Recommend a second opinion on this plan before we build.
README.md ↗ · Checked Oct 7, 2026skills/claudex-route/SKILL.md ↗ · Checked Oct 7, 2026
Author’s example

Harden an implementation plan

Uses claudex-loop

  • An implementation plan supplied in the conversation or available at a clearly identified project path
  • Both authenticated provider CLIs and Python 3.10+ for the cross-provider workflow
Use claudex-loop to harden this plan with independent review.
README.md ↗ · Checked Oct 7, 2026.codex-plugin/plugin.json ↗ · Checked Oct 7, 2026skills/claudex-loop/SKILL.md ↗ · Checked Oct 7, 2026
Suggested first task

Implement and independently inspect a work order

Uses codex-build

  • Path to a concrete work-order or specification file
  • An agreed proof command
  • Any explicit model, effort, log-path, or fix-round settings
Use codex-build to implement the concrete work order at <spec-path>, verify it with <proof-command>, and have Claude independently inspect the final changes. Replace <spec-path> with the actual specification file path and <proof-command> with the agreed command that verifies the deliverable.
README.md ↗ · Checked Oct 7, 2026skills/codex-build/SKILL.md ↗ · Checked Oct 7, 2026