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/CronusL-1141/AI-companynpx agentmods add skills/cronusl-1141/ai-company/os-releaseWrote 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/skills/cronusl-1141/ai-company/os-release)<a href="https://agentmods.dev/skills/cronusl-1141/ai-company/os-release"><img src="https://agentmods.dev/badge/skills/cronusl-1141/ai-company/os-release/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/skills/cronusl-1141/ai-company/os-release"><img src="https://agentmods.dev/badge/skills/cronusl-1141/ai-company/os-release.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 | $0.00094 | $0.05342 |
| Opus 5.5 | $0.00038 | $0.02137 |
| Sonnet 5.5 | $0.00019 | $0.01068 |
| Haiku 4.5 | $0.00009 | $0.00534 |
Grade C, and why
os-release scanned grade C with 1 finding 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf plugin/dashboard-dist How it starts
The opening of the file, as written. The whole thing — 319 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OS Release — 发版清单
为什么有这份清单
机检管得住的部分不用你记,跑一条命令即可。这份清单只管机检管不住的部分,以及 谁按哪个按钮。
历史上同型漏项两次:tag 推了、GitHub Release 条目没建,访客主页的 latest 徽章 因此卡在旧版本两周。所以别把"tag 推完"当成发版完成——第 10 步才是。
执行分工(硬约束:发布流水线必须可中断)
| 步骤 | 谁执行 |
|---|---|
| 1–6 准备与校验 | Leader 全权 |
| 7 commit · 8 tag | Leader 执行,但先把 git diff --stat 与拟好的 message 交用户批准 |
| 9 双仓推送 · 10 建 Release 条目 | 用户执行。Leader 只把命令准备好打印出来 |
| 11 事后核对 | Leader(只读) |
不要代替用户 push / publish,也不要在未批准前 commit。
1. 预检全绿
bash scripts/preflight.sh # 发版必须跑全量,不要 --fast
五道门禁:ruff check src/ tests/ → dashboard npm run lint → 前端实时事件回归
(npm --prefix dashboard test)→ check_invariants.sh → pytest tests/unit/。
机检条目数以脚本输出为准,别在别处写死范围。
preflight 不含、发版仍要跑的两项:pytest tests/integration/ 全过;
uvx --from build pyproject-build 产出 .whl 与 .tar.gz 无错(不往系统安装 build;
已有 build 的环境也可用 python -m build),再运行
python3 scripts/check_sdist_contents.py dist/ai_team_os-*.tar.gz 检查实际归档成员。
它拒绝本地 Python 环境、node_modules 与开发缓存/构建产物,并确认安装、源码、测试及
plugin/dashboard-dist 关键分发文件仍在,同时输出压缩字节数与成员数供核对;构建成功或
体积正常均不能代替内容门禁。
失败长这样:每项后面跟 ✗ <门禁名> 失败,末尾 ✗ 预检未通过 — 修复后再 push。
退出码非 0。
两处既定豁免:
- 未构建
dashboard/dist的环境里 I3 输出⚠️ dist 目录缺失(未构建环境可忽略)——警告不拦。但发版必须构建(见第 4 步),所以发版时这条不该还是警告。 - I16 在 execpolicy rules 或 Codex 二进制缺席时输出
⚠️ [I16] Codex execpolicy(...)——本期是占位状态,warn 不拦。它转成 ok 的那一天要留意:那意味着这台机器上真的跑了 一次宿主校验,结论才有分量。
用户生命周期验收(每次发版都做):新装一次、从上一个已发布版本更新一次,分别确认 更新后的代码、已安装的适配器、配置与运行中的服务实际生效;写明哪些变化自动生效、哪些 要重启或重新授信、哪些要用户显式启用。开发机上已应用、源码测试通过,都代替不了这一步。
Codex 装机面的严格对账也在这一步:每轮安装(新装、更新)完成后、打 tag 之前,跑
python3 scripts/check_codex_installed_hooks.py # 验收用的是隔离 Codex 目录时加 --codex-home <该目录>
退出码必须为 0。漂移或仅装侧即退出码 1,先查清是没装上还是装侧被改过,不许带红打 tag。
check_invariants 里的 I28 只警告(用户 2026-10-16 裁定,任务 00a77472:提交前工作区必然
领先于安装),这里是它的严格版。输出 [WARN] 本机没有 Codex 侧已安装副本 表示这台机器没装
Codex 适配器,验收记录里要写明「未验 Codex 装机面」,不能当作通过。
2. 版本锁步七处
I2 机检的五处(漂移即红,❌ [I2] 版本号漂移: ...):
pyproject.tomlsrc/aiteam/__init__.pyplugin/.claude-plugin/plugin.jsonplugin/.claude-plugin/marketplace.json.claude-plugin/marketplace.json
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 Changed · +18 lines scan D → C b271d2bfad0d
- 1mo ago Changed · +20 lines scan C → D 067de5cc895c
- 1mo ago Changed 42a2388ed88e
- 1mo ago Changed · +23 lines f61ed77c5e70
- 1mo ago Changed · -25 lines 44254904b550
- 1mo ago Changed · +10 lines 9180f50923fe
- 1mo ago Changed · +32 lines 7cbdfae86f77
- 2mo ago First seen · 241 lines · 94 tokens per session scan C c881629f9891
os-release is a skill published in the GitHub repository CronusL-1141/AI-company (374 stars, last pushed 10d ago), licensed MIT. It adds 94 tokens to every session and 5,342 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
os-release compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| os-releaseCronusL-1141/AI-company | Skill | 374 | 10d ago | MIT | C | 94 | Claude Code |
| agent-framework-py-releasemicrosoft/agent-framework | Skill | 15k | yesterday | MIT | A | 102 | Claude Code, Codex |
| pack-submitaeonfun/aeon | Skill | 788 | yesterday | MIT | A | 27 | Claude Code, Codex |
| shiplogaeonfun/aeon | Skill | 788 | yesterday | MIT | A | 46 | Claude Code, Codex |
| submit-hookaeonfun/aeon | Skill | 788 | yesterday | MIT | A | 39 | Claude Code, Codex |
| changelogcodeaholicguy/ai-devkit | Skill | 1.7k | 2d ago | Apache-2.0 | A | 49 | Claude Code, Codex |
| release-cutBlackBeltTechnology/pi-agent-dashboard | Skill | 341 | 2d ago | MIT | A | 93 | Claude Code, Codex |