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.
npx agentmods add instructions/wangnov/codex-app-manager/claude-mdgit clone --depth 1 https://github.com/Wangnov/Codex-App-ManagerWrote 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/instructions/wangnov/codex-app-manager/claude-md)<a href="https://agentmods.dev/instructions/wangnov/codex-app-manager/claude-md"><img src="https://agentmods.dev/badge/instructions/wangnov/codex-app-manager/claude-md.svg" alt="Measured on agentmods" 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 | $0.01026 | $0.01026 |
| Opus 5 | $0.00513 | $0.00513 |
| Sonnet 5 | $0.00205 | $0.00205 |
| Haiku 4.5 | $0.00103 | $0.00103 |
Grade A, and why
Codex-App-Manager CLAUDE.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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — Agent 作业指南
Codex App Manager:官方 OpenAI Codex 桌面应用的安装/更新/卸载管理器(Tauri v2)。
前端 src/,Rust 引擎 crates/ + src-tauri/,分发 Worker cloudflare/,官网 website/。
质量门与合并
- main 受保护:禁 force push、必须走 PR,必需检查 = Frontend + Rust (macos/windows)。
- 实质代码改动的收尾链路:
codex review --uncommitted(或--base main)迭代到无意见 → PR →gh pr merge --squash。 - 提交信息用英文 conventional 风格(
feat:/fix:/docs:/chore(release):)。
发版流程(tag 驱动)
- bump 版本号,5 个文件 6 处:
package.json、package-lock.json(顶层 +packages[""]两处)、src-tauri/tauri.conf.json、src-tauri/Cargo.toml、src-tauri/Cargo.lock(只改codex-app-manager那个[[package]]块——⚠️ lock 里winapi-util等依赖也是0.1.x, 严禁全局替换)。 - 同一个发版 PR 里写 release note:新增
docs/releases/v<X.Y.Z>.md。- 结构与双语风格照
docs/releases/TEMPLATE.md(banner 头图 → 一句话双语主旨 → ✨ 亮点 / 🐛 修复 双语成对行 → 安装升级表)。 - 素材来源:
git log <上个tag>..HEAD+ 读关键 diff 还原细节(squash 提交一句话 背后常有数百行变更);只写用户可感知的变化,内部重构不写。 - 事实核查:渠道、数字必须核实——例:winget(
Wangnov.CodexAppManager)在 microsoft/winget-pkgs 合并前不存在,不要写winget install。 - 安装表、镜像直链、更新器工件说明是模板固化内容,不要改写;Full Changelog 由 release.yml 自动追加,不要手写。
- 结构与双语风格照
- PR 标题
chore(release): bump version to <X.Y.Z>。bump PR 会额外触发非必需的nsis检查(~8min,期间 PR 状态 UNSTABLE)——等它也绿再合并。 - squash 合并后,在 main HEAD 打 annotated tag
v<X.Y.Z>并 push。release.yml自动:四平台构建(macOS arm64/x64 + Windows x64/ARM64)→ 签名/公证 → R2 + IHEP 镜像同步(失败会阻断发布)→ GitHub Release(正文取docs/releases/<tag>.md,缺失回退FALLBACK.md并告警,同时 追加自动 What's Changed)。 - 发布后会 dispatch
winget.yml;Wangnov.CodexAppManager已在 winget-pkgs 中可用, workflow 失败通常需要处理生成的提交/PR。README 下载链接是/manager/latest/*永久直链, 发版通常无需改 README。
website/ 子工程(官网,codexapp.agentsmirror.com)
- 独立 npm 工程,不要在仓库根目录为它装依赖(会污染主工程 package.json)。
- 文案唯一来源
website/src/locales/{zh,en}.ts;改文案后重跑npm run fonts(中文显示字体按用字子集化,缺字会回退系统字体)。 - 素材管线:
assets/raw/(git-ignored,AI 生成)→npm run images→public/img/。 - 部署:
cd website && npm run build && npx wrangler deploy。zone 路由:/manager/*→ 本仓库下载路由器,/latest/*→ mirror 仓库,/*→ 官网,互不抢路。 - README banner(
assets/banner.svg)由node website/scripts/readme-banner.mjs再生;官网视觉资产更新后记得重新生成。
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.
- 4d ago First seen · 47 lines · 1,026 tokens per session scan A b4435479ab9c
Codex-App-Manager CLAUDE.md is an instructions file published in the GitHub repository Wangnov/Codex-App-Manager (378 stars, last pushed 6d ago), licensed MIT. It adds 1,026 tokens to every session, about $0.0051 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.
Other instructions, from other repositories
moonproxy-desktop AGENTS.md
Instructions for MoonProxyHQ/moonproxy-desktop, covering moonproxy, 功能, 技术栈, 目录结构 and 配置与数据存储.
fyagent AGENTS.md
AGENTS.md instructions for fy-agent/fyagent: These instructions are for AI assistants working in this project.
workhub CLAUDE.md
Claude Code instructions for atman-33/workhub, covering workhub, commands, in src-tauri/, architecture and data model: the vault is the source of truth.
WeftCut AGENTS.md
Instructions for WeftCut/WeftCut, covering weftcut — agent instructions, agent skills, issue tracker, domain docs and testing.
desk_tidy_sticky AGENTS.md
AGENTS.md instructions for sqmw/desk_tidy_sticky, covering project notes and workspace paths.
ghost CLAUDE.md
Claude Code instructions for ghostapp-ai/ghost, covering ghost — agent development instructions, project overview, critical rules, 1. never break privacy and 2. performance is non-negotiable.