project-workflow AGENTS.md

A project-level instruction file for AI coding agents working in the project-workflow repository. It explains the repository structure, documentation, development rules, and boundaries.

In plain words
What is it for?
Use it to guide feature development across the project’s documentation, templates, and Claude Code and Codex adapters.
Why use it?
It gives agents shared project context and rules so their changes follow the repository’s workflow and limits.

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/shrekshrek/project-workflow/agents-md
Clone the repo
git clone --depth 1 https://github.com/shrekshrek/project-workflow

Made for: Codex, OpenCode.

Per session 2,517 This file is loaded in full into every session.
When invoked 2,517 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 $0.02517 $0.02517
Opus 5 $0.01259 $0.01259
Sonnet 5 $0.00503 $0.00503
Haiku 4.5 $0.00252 $0.00252

Measured 2d ago against content hash 46e9db8d364f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

project-workflow 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 · 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.

AGENTS.md

Project-level instructions for any AI coding agent (Claude Code, Codex, OpenCode, Cursor). Claude Code reads CLAUDE.md which is @AGENTS.md.

仓库性质

project-workflow v3 —— conversation-and-evidence-led feature development workflow + runtime adapters。

四层资产(本仓库):

  1. docs/ —— 方法论文档(四个运行阶段、4 支柱、canonical actions、按需单份 spec、gotchas 示范 ledger)
  2. template/ —— starter scaffold(方法论核心 + Claude compatibility assets,语言中立但非 tool-empty)
  3. adapters/claude/ —— Claude Code-native skills / named agents / manifest
  4. adapters/codex/ —— Codex-native skills / manifest;两端安装包由构建脚本生成

遵循 project-workflow 方法论的一个具体项目示例:shrekshrek/full-stack-scaffolding-fastapi-nuxt4 —— FastAPI + Nuxt 4 全栈脚手架,公开。仅作 example-of-one,方法论自身的论据不依赖它。

文档索引

README.md              v3 总览 + 安装 + skill 清单
.claude-plugin/        Claude marketplace 注册(指向 plugin-dist)
.agents/plugins/       Codex marketplace 注册(指向 plugin-dist)
adapters/
├── claude/
│   ├── .claude-plugin/plugin.json
│   ├── agents/        6 个 Claude named-agent thin adapters
│   └── skills/        Claude Code skills(9 个,与 Codex 同一 action surface,无 helper 层)
└── codex/
    ├── .codex-plugin/plugin.json
    └── skills/        Codex-native skills(9 个,可调度时强制通用 subagent + 有据 fallback)

adapters/claude/skills/
├── project-init/         /project-workflow:project-init        (P0 baseline-compatible / already-initialized target)
├── project-personalize/  /project-workflow:project-personalize (P0 project evidence / partial-custom baseline)
├── feature-init/         /project-workflow:feature-init        (P2 沟通、试验与按需记录)
├── spec-quality-check/   /project-workflow:spec-quality-check  (P2 语义就绪与需求对账)
├── spec-revise/          /project-workflow:spec-revise         (P2 mid-impl 修订,§3.5/§2.6)
├── feature-done/         /project-workflow:feature-done        (P2 端点:L1/L2/L3/current-truth/proof 一体,幂等)
├── feature-archive/      /project-workflow:feature-archive     (交付后生命周期收尾:current truth + 老 spec 标记)
├── spec-reconcile/       /project-workflow:spec-reconcile      (多 spec 漂移诊断 + 状态修正)
└── agents-md-revise/     /project-workflow:agents-md-revise    (P4 主动 refresh A 类约定)

adapters/claude/agents/ Claude Code sub-agent adapters(thin wrappers over docs/reviewers/)
├── agents-md-reviewer.md            L2 AGENTS.md 合规 review(by /feature-done L2 层)
├── spec-reviewer.md                 L3 spec.md 合规 review(by /feature-done L3 层)
├── spec-quality-reviewer.md         需求对账 + spec 质量与条件式架构充分性复核(by /spec-quality-check)
├── tech-researcher.md               技术选型调研(by /project-personalize;spec 填写时仅作可选研究,不为覆盖率强制调用)
├── codebase-explorer.md             非平凡 codebase 结构扫描(by /project-personalize 按需调用)
└── decision-completeness-auditor.md plant 决策追溯审计(by /project-personalize / /feature-init / /spec-revise / /agents-md-revise,Preview Gate 之前)— 实施 workflow.md §1.12
docs/                  方法论文档
├── actions/           workflow action canonical specs
├── reviewers/         reviewer/auditor/researcher canonical specs
├── architecture-design.md  architecture change 的按需设计指南
├── workflow.md        ⭐ 四个运行阶段 + 4 支柱(核心)
├── cross-tool-methodology.md  core vs runtime adapter 边界
├── gotchas.md         example-of-one gotchas 证据库(示范短版)
├── spec-driven.md     沟通、记录与验证详解
└── tooling.md         工具链对比
template/              starter scaffold(core files + Claude compatibility assets)
scripts/
├── check-all.cjs  run the complete local validation suite and summarize failures
├── build-plugin-packages.cjs  generate/check Claude + Codex self-contained packages
├── check-adapter-parity.js  check 9+9 action parity + runtime isolation
├── check-template-contracts.js  check baseline safety + Claude rule frontmatter
├── check-reviewer-fixtures.cjs  check reviewer smoke fixture inputs + verdict truth tables
├── check-feature-init-fixtures.cjs  check/grade feature-init behavior scenario matrix
├── materialize-feature-artifact.cjs  atomically create a no-clobber feature artifact from the single spec template
├── materialize-project-baseline.cjs  stage/apply P0 baseline (used by project-init / project-personalize)
├── relocate-markdown-links.cjs  preserve local links after feature archive moves
├── check-lifecycle-links.cjs  regression-check archive link relocation
├── check-markdown-links.cjs  verify local Markdown destinations across source + runtime adapters + release docs
└── check-workflow-contracts.cjs  check conversation / records / retrofit / verdict / neutral-template semantics

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. 2d ago First seen · 103 lines · 2,517 tokens per session scan A 46e9db8d364f

Subscribe to this mod's changes

project-workflow AGENTS.md is an instructions file published in the GitHub repository shrekshrek/project-workflow (2 stars, last pushed 2d ago), licensed MIT. It adds 2,517 tokens to every session, about $0.0126 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.