Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/treylom/tofu-atnpx agentmods add commands/treylom/tofu-at/tofu-at-codexWrote 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.
[](https://agentmods.dev/commands/treylom/tofu-at/tofu-at-codex)<a href="https://agentmods.dev/commands/treylom/tofu-at/tofu-at-codex"><img src="https://agentmods.dev/badge/commands/treylom/tofu-at/tofu-at-codex/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.
<a href="https://agentmods.dev/commands/treylom/tofu-at/tofu-at-codex"><img src="https://agentmods.dev/badge/commands/treylom/tofu-at/tofu-at-codex.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00027 | $0.10458 |
| Opus 5 | $0.00014 | $0.05229 |
| Sonnet 5 | $0.00005 | $0.02092 |
| Haiku 4.5 | $0.00003 | $0.01046 |
Grade B, and why
tofu-at-codex scanned grade B with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
| 1 | **tmux** | Agent Teams Split Pane | `sudo apt install tmux` (Linux) / `brew install tmux` (macOS) | Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
HEALTH=$(curl -s -o /dev/null -w '%{http_code}' \ How it starts
The opening of the file, as written. The whole thing — 916 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/tofu-at-codex — Codex Hybrid Team v2
Leader = Opus (Anthropic Direct) Workers = Sonnet/Opus (Anthropic Direct) DA = Codex adversarial review (
codex exec) PR Review =codex review --base(선택)기존
/tofu-at워크플로우를 그대로 사용하되, DA 역할을 Codex adversarial review로 실행하는 하이브리드 모드입니다.v2 변경사항 (2026-03-31 autoresearch 실험 기반):
- CLIProxyAPI 제거 — Workers는 Anthropic Direct로 실행
- DA를
codex execadversarial review로 교체 (97초, score 11.0)codex review --base를 PR/diff 리뷰 옵션으로 추가- 의존성 3개 → 2개로 축소 (CLIProxyAPI, OAuth 불필요)
$ARGUMENTS
Quick Start (설치 가이드)
원클릭 환경 체크
bash .claude/scripts/setup-tofu-at-codex.sh
의존성 목록
| # | 의존성 | 용도 | 설치 명령어 |
|---|---|---|---|
| 1 | tmux | Agent Teams Split Pane | sudo apt install tmux (Linux) / brew install tmux (macOS) |
| 2 | Codex CLI | DA adversarial review + PR review | npm install -g @openai/codex && codex login |
| 3 | Claude Code | Agent Teams 기반 | docs.anthropic.com |
v2에서 제거된 의존성: CLIProxyAPI, OAuth 토큰 (더 이상 불필요)
Codex Plugin (선택)
codex review 기능을 Claude Code 내에서 slash command로 사용하려면:
/plugin marketplace add openai/codex-plugin-cc
/plugin install codex@openai-codex
/reload-plugins
실행 방법
# 1. tmux 세션 시작 (이미 tmux 안이면 스킵)
tmux new -s claude
# 2. Claude Code 실행
claude
# 3. /tofu-at-codex 실행
/tofu-at-codex
v3.0 핵심 아키텍처 (autoresearch 17라운드 기반, R0-R17):
| 역할 | 모델 | 실행 방식 |
|---|---|---|
| Leader | Opus 4.6 [1M] | Anthropic Direct |
| Workers (비코딩) | Sonnet 4.6 | Anthropic Direct (팀원 스폰) |
| Workers (코딩) | GPT-5.4 | codex exec --full-auto (Lead 직접 호출) |
| DA | GPT-5.4 | adversarial-review --scope working-tree (Lead 직접 호출) |
| PR Review | GPT-5.4 | codex review --base (선택) |
작업 유형별 Codex 라우팅 (CRITICAL):
| 작업 유형 | 최적 전략 | Score | 근거 |
|---|---|---|---|
| 코딩 (버그 수정, 구현, 리팩토링) | codex exec 직접 |
11.0 | R4: 67초, 팀 대비 4.7x 빠름, 품질 동일 |
| 비코딩 (리서치, 분석, 문서화) | Anthropic Workers 스폰 | 10.0 | R6: 품질 8/10, codex보다 상세 |
| DA (리뷰, 반론 검증) | adversarial-review --scope wt |
11.0 | R14: R2 동점, prompt_effort 3→0 |
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.
- 9d ago First seen · 916 lines · 27 tokens per session scan B 82d5380e12a8
tofu-at-codex is a command published in the GitHub repository treylom/tofu-at (56 stars, last pushed 1mo ago), licensed MIT. It adds 27 tokens to every session and 10,458 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 2 findings (asks for root, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
sdd-init
Initialize SDD context — detects project stack and bootstraps persistence backend.
review-branch
Review the current branch's diff against base by dispatching atomic-reviewer. No orchestration loop, no spec required — pre-flight before /commit pr or /commit merge.
init
Install the formatters this repository needs, with every command visible before it runs.
repo-audit
Audit a codebase (local or remote GitHub/GitLab) against architecture principles and requirements, surfacing drift, risk, and missing decisions.
argos
A command for checking whether an implementation matches its design deliverables. Its Korean description compares the work to the design as part of a completion inspection.
security-review
AI-powered security review of the current git diff (or specified paths). Dispatches the security-reviewer agent and prints findings grouped by severity.