dual-ai-paper-coach: Command for Claude Code

.claude/commands/codex-review.md

codex-review is a command for Claude Code, Codex from dengxu11111/dual-ai-paper-coach. It costs 18 tokens per session (1,247 once invoked), scanned A, original, MIT.

A command that starts the Codex command-line tool as an independent reviewer for a scientific paper draft.

In plain words
What is it for?
Use it to review a draft paper, check quantitative claims when an analysis-results file exists, and save a structured peer-review report for the next review round.
Why use it?
It provides a separate review of the manuscript and can compare numerical claims with the project’s analysis results instead of relying on the main writing assistant alone.

Command for Claude CodeCodex

Written for Claude Code and Codex: installed under .claude/, but also runs codex exec. Also seen: mentions AGENTS.md; mentions Codex.

This is dengxu11111/dual-ai-paper-coach's own configuration. It tells Claude Code and Codex how to work on dual-ai-paper-coach itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything dual-ai-paper-coach configures →

Reuse

Borrowing it

Nothing to install: this file belongs to dengxu11111/dual-ai-paper-coach. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/dengxu11111/dual-ai-paper-coach/main/.claude/commands/codex-review.md
Clone the repo
git clone --depth 1 https://github.com/dengxu11111/dual-ai-paper-coach

Made for: Claude Code, Codex.

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 codex-review

README.md
[![agentmods](https://agentmods.dev/badge/commands/dengxu11111/dual-ai-paper-coach/codex-review/github.svg)](https://agentmods.dev/commands/dengxu11111/dual-ai-paper-coach/codex-review)
Your own site
<a href="https://agentmods.dev/commands/dengxu11111/dual-ai-paper-coach/codex-review"><img src="https://agentmods.dev/badge/commands/dengxu11111/dual-ai-paper-coach/codex-review/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for codex-review

Your own site · 80×15
<a href="https://agentmods.dev/commands/dengxu11111/dual-ai-paper-coach/codex-review"><img src="https://agentmods.dev/badge/commands/dengxu11111/dual-ai-paper-coach/codex-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 18 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,247 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00018 $0.01247
Opus 5 $0.00009 $0.00624
Sonnet 5 $0.00004 $0.00249
Haiku 4.5 $0.00002 $0.00125

Measured 12d ago against content hash fd65c6487813, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

codex-review 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 12d 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.

.claude/commands/codex-review.md · 103 lines

How it starts

The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/codex-review — Codex 独立审稿

重要:这是双 AI 流程的核心独立性边界。你(Claude)不审这份稿子。你只是触发 Codex 子进程,然后读 Codex 写出的报告。

工作流

用户跑 /codex-review
    ↓
你(Claude)通过 Bash 调用 codex exec
    ↓
Codex CLI 启动子进程(独立环境)
    ↓
Codex 读 AGENTS.md + prompts/reviewer_system.md + my-paper/draft.md + my-paper/data/analysis_results.json(如果存在)
    ↓
Codex 输出 review 到 my-paper/review_round_N.md
    ↓
你读这份 review 文件,给用户预览前 30 行

步骤

步骤 1:确定轮次号

my-paper/ 下已经存在的 review_round_*.md 文件,找到最大 N,新一轮为 N+1。

步骤 2:构造 Codex 调用

用 Bash 工具跑:

codex exec --skip-git-repo-check --sandbox workspace-write \
  -c 'model_reasoning_effort="xhigh"' \
  "Read AGENTS.md, prompts/reviewer_system.md, prompts/review_paper.md, my-paper/draft.md, and if present my-paper/data/analysis_results.json. The {manuscript} placeholder in review_paper.md should be replaced with the contents of draft.md. If analysis_results.json is present, verify every quantitative manuscript claim against it and flag any mismatch or unverifiable number as a Critical Issue. If it is absent, state that numerical verification was unavailable. Produce a structured peer review per the format in reviewer_system.md. Write your output to my-paper/review_round_<N>.md (replace <N> with the round number ${ROUND}). Do not modify any other files."

${ROUND} 替换成步骤 1 算出的轮次号。

关键参数

  • --skip-git-repo-check:避免 Codex 卡在 git repo 检查
  • --sandbox workspace-write必须——默认沙盒只读,Codex 写不进 review_round_N.md
  • -c 'model_reasoning_effort="xhigh"':审稿要用 xhigh,默认 medium 不够严
  • 不传 --model,让 Codex 用它的默认模型
  • 显式让 Codex 读 my-paper/data/analysis_results.json(如存在):否则它只能审文字,不能兑现"数字必须能对上数据"这条红线

步骤 3:等 Codex 完成

Codex 子进程通常运行 1-4 分钟(视论文长度、网络和账号状态);慢时可能到 5-10 分钟。需要快速对比可参考预跑好的 examples/global-aerosol-2010-2024/review_round_1.md

步骤 4:读结果并预览

Codex 跑完后:

  1. 检查 my-paper/review_round_<N>.md 是否存在
  2. 不存在 → 报错"Codex did not produce review file. Check codex --help and Codex login status."
  3. 存在 → 读它,提取 ## Score 那一行的分数
  4. 用 markdown 预览给用户看
  5. 同时把这次运行的元数据写到 my-paper/runs/<ts>/codex_review_meta.json
    • 轮次号
    • Codex 版本(通过 codex --version
    • 论文字数
    • Score
    • Verdict

Read the full file on GitHub · 103 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. 12d ago First seen · 103 lines · 18 tokens per session scan A fd65c6487813

Subscribe to this mod's changes

codex-review is a command published in the GitHub repository dengxu11111/dual-ai-paper-coach (12 stars, last pushed 3mo ago), licensed MIT. It adds 18 tokens to every session and 1,247 once invoked, about $0.0001 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 2026-08-30.