codex-oi AGENTS.md

A project instruction file explaining how coding agents can use codex-oi, a terminal tool that asks another AI for a code review or planning opinion. AGENTS.md is a shared convention supported by several coding tools.

In plain words
What is it for?
Copy the instructions into a project's AGENTS.md file to guide second-opinion reviews, closeout checks, and plan reviews through codex-oi. It also explains how to choose a review mode and engine.
Why use it?
It gives agents clear rules for requesting a second review before a significant commit or when someone asks for an audit. This helps expose issues that one review might miss.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/dtdkvn/codex-oi/agents-md
Clone the repo
git clone --depth 1 https://github.com/Dtdkvn/codex-oi

Made for: Codex, OpenCode.

Per session 1,510 This file is loaded in full into every session.
When invoked 1,510 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.01510 $0.01510
Opus 5 $0.00755 $0.00755
Sonnet 5 $0.00302 $0.00302
Haiku 4.5 $0.00151 $0.00151

Measured 2d ago against content hash 9ab79a052042, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

codex-oi AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 150 lines

How it starts

The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md — codex-oi

Cross-tool agent rules file. Recognized by Antigravity (v1.20.3+), Cursor, Codex CLI, Claude Code, and any tool that adopts the AGENTS.md convention. Use this file to teach your coding agent when and how to invoke codex-oi for a second opinion.

This repo IS codex-oi. The note below is meant to be copied into the AGENTS.md of any project that wants the agent to use codex-oi as a second-opinion engine. For working on this repo itself, see CLAUDE.md (if present) and SKILL.md.


Copy-paste block — drop into your project's AGENTS.md

## Second-opinion reviews via codex-oi

This repo has `codex-oi` available — a Codex CLI bridge that gives a
second AI opinion on code without leaving the terminal. Use it
proactively in these situations:

1. **Before committing a non-trivial diff** — run a closeout check:
   ```bash
   ~/.codex-oi/scripts/codex-oi.sh closeout
  1. When the user asks for "a second opinion", "audit", or "review by another model" — pick the matching mode:

    ~/.codex-oi/scripts/codex-oi.sh review <file-or-folder> "<focus>"
    ~/.codex-oi/scripts/codex-oi.sh plan <plan.md>
    ~/.codex-oi/scripts/codex-oi.sh audit       # full project sweep, cost-warned
    
  2. After landing a non-trivial commit — retrospective:

    ~/.codex-oi/scripts/codex-oi.sh recommit HEAD
    

Treat the output as advisory only. Verify every finding against the real code path. Reject speculative or out-of-scope claims with a one-line reason. Do not auto-apply fixes.

Windows-native (no WSL): use scripts\codex-oi.ps1 with the same modes.


---

## Pluggable second-opinion engine (`--engine` flag)

Default engine is OpenAI Codex. Other engines available via `--engine`:

```bash
~/.codex-oi/scripts/codex-oi.sh review file.py "focus" --engine gemini

Currently shipped adapters (Bash dispatcher only; PowerShell mirror TODO):

Read the full file on GitHub · 150 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 150 lines · 1,510 tokens per session scan A 9ab79a052042

Subscribe to this mod's changes

codex-oi AGENTS.md is an instructions file published in the GitHub repository Dtdkvn/codex-oi (2 stars, last pushed 3mo ago), licensed MIT. It adds 1,510 tokens to every session, about $0.0076 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.