@steipete: Wrote a skill that runs codex /review in a loop until there's no booboos anymore. Caveat: It won't fix system architect…
Summary
Peter (@steipete) created a skill that runs Codex's /review command in a loop to fix code issues, but notes it doesn't fix system architecture. The associated GitHub repo 'agent-scripts' contains shared agent instructions, skills, and helper scripts for local workspaces.
View Cached Full Text
Cached at: 05/15/26, 03:03 PM
Wrote a skill that runs codex /review in a loop until there’s no booboos anymore.
Caveat: It won’t fix system architecture for ya, so you still need BRAIN as master model. https://t.co/0Z6iJnCqCX
steipete/agent-scripts
Source: https://github.com/steipete/agent-scripts
Agent Scripts
Shared agent instructions, skills, and small portable helpers for Peter’s local workspaces.
This repo is the canonical place for:
AGENTS.MD: shared hard rules for Codex/Claude-style agentsskills/: reusable workflow skills, including repo-owned skills exposed by symlinkscripts/: dependency-light helpers used across projectshooks/: local guardrails such as skill validation
Skills
Skills are the main routing layer. Each skills/<name>/SKILL.md has YAML front matter:
---
name: skill-name
description: "Short generic trigger phrase."
---
Rules:
- Keep descriptions short and generic; optimize for routing, not documentation.
- Keep skill bodies terse and operational.
- Prefer helper scripts under
skills/<name>/scripts/when a workflow has repeatable commands. - Validate after edits:
scripts/validate-skills. - Quote
descriptionin front matter.
Global discovery usually points here:
~/.codex/skills -> ~/Projects/agent-scripts/skills~/.claude/skills -> ~/Projects/agent-scripts/skills
Shared skills live as real folders in skills/. Repo-owned skills stay canonical in their repo and are exposed here with tracked relative symlinks, for example:
skills/discrawl -> ../../discrawl/.agents/skills/discrawl
Current symlinked repo-owned skills include birdclaw, discrawl, gog, imsg, slacrawl, wacli, and wacrawl.
Agent Instructions
Shared hard rules live in AGENTS.MD.
Global setup:
~/.codex/AGENTS.md -> ~/Projects/agent-scripts/AGENTS.MD~/.claude/CLAUDE.md -> ~/Projects/agent-scripts/AGENTS.MD~/.claude/AGENTS.md -> ~/Projects/agent-scripts/AGENTS.MD
Downstream repos should use a pointer-style AGENTS.MD:
READ ~/Projects/agent-scripts/AGENTS.MD BEFORE ANYTHING (skip if missing).
Repo-specific rules go below that pointer. Do not copy the shared blocks into downstream repos.
Helpers
scripts/committer
- Stages exactly the listed files.
- Enforces a non-empty commit message.
- Runs skill validation before committing.
scripts/validate-skills
- Checks every
skills/*/SKILL.md. - Verifies YAML front matter plus required
nameanddescription. - Enable as a local hook with
git config core.hooksPath hooks.
scripts/docs-list.ts
- Walks
docs/. - Enforces
summaryandread_whenfront matter. - Prints onboarding summaries for repos that wire it in.
scripts/browser-tools.ts
- Standalone Chrome DevTools helper.
- Common commands:
start --profile,nav <url>,eval '<js>',screenshot,search --content "<query>",content <url>,inspect,kill --all --force. - Build optional binary with
bun build scripts/browser-tools.ts --compile --target bun --outfile bin/browser-tools.
Syncing
Treat this repo as canonical for shared agent rules and portable helper scripts.
When syncing downstream repos:
- Pull latest here first.
- Ensure each target repo starts with the pointer-style
AGENTS.MD. - Preserve repo-local rules below the pointer.
- Copy helper changes both directions only when the helper is meant to stay byte-identical.
- Keep scripts dependency-free and portable; no repo-specific imports or path aliases.
For submodules, repeat the pointer check inside each subrepo, push those changes, then bump submodule SHAs in the parent repo.
Similar Articles
Building Self-Repairing Agent Loops (39 minute read)
This article presents a method for building self-repairing agent loops using OpenAI's Codex, where agents review, repair, and validate outputs iteratively, with a worked example of fixing stale API documentation.
@Saboo_Shubham_: Codex /goal builds it. Claude Code /goal review and refines it. Hermes /goal manages the orchestration and handoff. All…
A demonstration of a multi-agent workflow where Codex builds code, Claude Code reviews it, and Hermes manages orchestration, all tracked on a Kanban board.
I built a local CLI for Claude Code, Codex, and Gemini to review each other’s GitHub PRs usign existing auth
The author introduces `coding-review-agent-loop`, an open-source local CLI that orchestrates multiple coding agents (Claude Code, Codex, Gemini) to review each other's GitHub PRs using existing local authentication, avoiding additional API costs.
@dr_cintas: Obsidian CEO personally wrote the official Agent Skills for his own app These are 5 skills that fix every layer agents …
The Obsidian CEO released five MIT-licensed official Agent Skills designed to fix common issues with AI agents interacting with the app, supporting tools like Claude Code and Codex CLI.
mattpocock/skills
This open-source repository provides a composable set of AI agent skills and prompts designed to improve alignment, reduce verbosity, and optimize workflows for coding assistants like Claude Code and Codex.