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 skills add 0fuk/OpenACCP --skill moniter-ongit clone --depth 1 https://github.com/0fuk/OpenACCPWrote 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/0fuk/openaccp/moniter-on)<a href="https://agentmods.dev/skills/0fuk/openaccp/moniter-on"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/moniter-on/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/0fuk/openaccp/moniter-on"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/moniter-on.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 | $0.00242 | $0.04973 |
| Opus 5.5 | $0.00097 | $0.01989 |
| Sonnet 5.5 | $0.00048 | $0.00995 |
| Haiku 4.5 | $0.00024 | $0.00497 |
Grade B, and why
moniter-on scanned grade B 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 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
在 `~/.claude/settings.json`(或各值守项目的 `.claude/settings.json`)加: How it starts
The opening of the file, as written. The whole thing — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/moniter-on v3 — 无人值守可达性:强制重架 + 通道实测(发一个真事件看它能否回到会话)
存在理由:owner 不在时,worker 唯一合法的醒来方式是持久 Monitor 扫本地文件+文件信箱回信;任何依赖人工点 allow once 的交互工具(send_message/AskUserQuestion)都不能作为等待机制。本 skill 把"接上电"标准化、可验证化。
v4 变更说明(20260731,owner 亲令:"监控是loop,不要bash"——实现层整体换轨,本节覆盖下文一切 bash Monitor 做法)
**bash while-loop 型 Monitor 全面退役。**owner 两次实锤(20260730"不要bash,要loop,bash会掉"+20260731 本令):bash 常驻循环绑定会话进程,重载/压缩后必成残留进程或断链,v3 的"清残留进程+通道实测(发一个真事件看它能否回到会话)"是在给一个不该存在的实现方式打补丁。
v4 唯一合法实现=会话外部调度的 /loop 周期扫描(ScheduleWakeup):
- 文件信箱监控=每次 /loop tick 里用 find 按 mtime 多留 30 分钟的时间窗扫协调目录(窗口=上次 tick 时刻再往前多留 30 分钟重叠,宁重读勿漏读——20260731 两起漏收实锤都是窄窗边界:07:33 复位件、09:28-36 三张 C13 交付全漏);pattern 仍按 §2 从真实文件名倒推+回测。
- tick 节奏:活跃派单窗 20-30 分钟一 tick;低能耗值守窗 60 分钟;每次 tick 开头刷心跳文件(date 实钟)。
- 通道实测(发一个真事件看它能否回到会话)=v3 的 SELFTEST 文件改为"下一个 tick 真实扫到并处置"即闭环;不再依赖 40 秒轮询回报。
- 本 skill 被调用时:①TaskStop 一切在册 bash Monitor+按 v3 §1.2 清残留进程(清理历史存量)②确认 ScheduleWakeup 环在飞且 prompt 含宽窗扫描步骤,没有就当场排③写 SELFTEST 件候下一 tick 自我声明。
④(20260904 owner 实锤补)周期环唯一性:重架前先
CronList,把上一代留下的 session-only cron 逐条CronDelete,再ScheduleWakeup stop清掉旧动态环,然后只排一条;周期环只许有一个(ScheduleWakeup 或 cron 二选一,禁并存)。20260904 主编排两代 /moniter-on 各建一条 cron 未删旧+又叠 ScheduleWakeup=三个环同飞,owner 面板看见两个 compass loop 才发现——重架不幂等就是这样来的。 - 例外:确需秒级响应的短命定向监听(如等一次构建完成)用 Bash run_in_background 的 until 单发,完成即退出——禁 persistent while-true。
下文 v3 的 §1(清残留进程)、§2(pattern 倒推+回测)、§3.5(关闭铁律)、§5.5(被质疑即重架)仍然有效;§3 的 bash Monitor 模板与 §3.6 的 40 秒自我声明仅作历史参考,禁再使用。
v3 变更说明(20260721,必读——判据本身修正,不只是流程修正)
v2 的判据有一个实锤盲区:v2 用 标记文件 mtime 判活("60 秒内新鲜=Monitor 在跑,无需重架")。20260721 owner 在多个 thread 反复观察到:background tasks 面板空空如也、派单确实收不到,但模型每次都拿"标记文件 几秒前刚刷新"当证据回答"存活无需重架"。复盘定性:两边都是真的——
- Monitor 的 bash 循环是绑定会话进程的后台任务;会话重载/压缩重启/桌面端刷新后,平台对这个任务的登记没了(通知管道断了),但 OS 层的 bash 进程可能没死,它变成残留进程:继续每 20 秒 touch 标记文件、继续刷文件信箱心跳文件,唯独"发现新文件→通知会话"这条最关键的链路已经断了;
- 于是 标记文件 新鲜(残留进程在刷)+ 面板零 Running(平台不再登记)+ 派单永远收不到(事件无处投递)三者同时成立;
- 标记文件 mtime 只能证明"有个进程活着",永远不能证明"通知通道活着"。v2 让模型拿它当"无需重架"的依据,是判据错误,不是模型态度问题——但结果一样:owner 说挂了,模型"贴证据狡辩",实际 owner 每次都是对的。
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.
- 2d ago First seen · 198 lines · 242 tokens per session scan B 86b54462c001
moniter-on is a skill published in the GitHub repository 0fuk/OpenACCP (2 stars, last pushed 2d ago), licensed MIT. It adds 242 tokens to every session and 4,973 once invoked, about $0.0010 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-14.
Other skills, from other repositories
hive-mind-advanced
Queen-led collective-intelligence coordination with consensus mechanisms and persistent shared memory. Use when orchestrating a large swarm of agents under a lead coordinator, building consensus across agents, or maintaining durable cross-agent memory.
handoff-mesh
Built-in communication layer for multi-agent setups. Full-mesh handoff folders (any agent can hand off to any other), an hourly watcher that surfaces open handoffs, weekly archive cleanup, and a completion-reporting loop so every handoff is handled AND verified. The durable 'brain' counterpart to real-time channels…
tmux
Manage long-running shell sessions with tmux — start a detached session, run a long task, reattach later, capture output. Use when a task takes longer than a single tool call (build, test, log tail).
cog-braindump-capture
Capture raw thoughts with automatic domain classification and vault routing.
cog-knowledge-consolidation
Build structured knowledge frameworks from scattered vault notes with source attribution.
todos
Use when the user asks for a shared plan, a task needs at least three substantive execution steps, or a loose TODO is pending. Without an explicit plan request, not for one-shot answers or checks, or one- to two-step work that is not already in the list.