claude-codex-subagent AGENTS.md

claude-codex-subagent AGENTS.md is an instructions file for Codex, OpenCode from dwgx/claude-codex-subagent. It costs 960 tokens per session, scanned A, original, MIT.

Repository instructions for a Claude Code add-on that delegates expensive work to the local Codex command-line tool. They define the repository’s files, supported Codex version, editing rules, and validation guidance.

In plain words
What is it for?
Maintaining the delegation skill, dispatch script, documentation, and persona templates in that repository.
Why use it?
They give agents a shared set of rules for keeping the delegation workflow accurate and runnable.

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/dwgx/claude-codex-subagent/agents-md
Clone the repo
git clone --depth 1 https://github.com/dwgx/claude-codex-subagent

Made for: Codex, OpenCode.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for claude-codex-subagent AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/dwgx/claude-codex-subagent/agents-md.svg)](https://agentmods.dev/instructions/dwgx/claude-codex-subagent/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/dwgx/claude-codex-subagent/agents-md"><img src="https://agentmods.dev/badge/instructions/dwgx/claude-codex-subagent/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 960 This file is loaded in full into every session.
When invoked 960 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 $0.00960 $0.00960
Opus 5 $0.00480 $0.00480
Sonnet 5 $0.00192 $0.00192
Haiku 4.5 $0.00096 $0.00096

Measured 3d ago against content hash a6132f0ec7a4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

claude-codex-subagent 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 3d 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 · 92 lines

How it starts

The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.

claude-codex-subagent Agent Rules

This repository packages a Claude Code skill that teaches Claude to delegate expensive work to local codex exec. The job of every agent in this repo is to keep that delegation workflow accurate, current, and directly runnable.

Repository Role

  • skills/codex-subagent/SKILL.md is the source of truth for Claude behavior.
  • scripts/codex-dispatch.sh is the executable reference implementation.
  • docs/codex-dispatch-guide.md is the detailed parameter and workflow guide for both Claude and Codex.
  • docs/prompt-and-agent-patterns.md is the public guide for prompt, agent, skill, command, result-handling, and publication patterns.
  • personas/*.md are prompt templates. Their frontmatter must stay simple: persona, sandbox, effort, and when-to-use.

Codex CLI Baseline

Target Codex CLI 0.141.0 or newer.

  • Prefer --sandbox workspace-write for normal local repo work.
  • Do not introduce new --full-auto examples. It is a deprecated compatibility alias; scripts may accept full-auto as a user-facing alias only if they emit --sandbox workspace-write.
  • Use --dangerously-bypass-approvals-and-sandbox only for tasks that truly need shell network, cross-workspace access, global installs, or as a documented Windows sandbox fallback.
  • Use helper --search for current public web facts when shell network is not needed; direct codex exec uses --config web_search="live".
  • Use --json, -o/--output-last-message, and --output-schema when the result is consumed by automation.
  • Keep --skip-git-repo-check in all codex exec examples and wrappers.
  • Treat resume as a separate parameter surface: --cd, --profile, --add-dir, and non-bypass --sandbox are fresh-session context flags and should be rejected by the helper on --resume; --model, --config, --json, -o, --output-schema, --ephemeral, and explicit bypass are resume-time controls in Codex CLI 0.141.0.

Editing Rules

Read the full file on GitHub · 92 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. 3d ago First seen · 92 lines · 960 tokens per session scan A a6132f0ec7a4

Subscribe to this mod's changes

claude-codex-subagent AGENTS.md is an instructions file published in the GitHub repository dwgx/claude-codex-subagent (3 stars, last pushed 1mo ago), licensed MIT. It adds 960 tokens to every session, about $0.0048 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.