A navigation and progress audit for agent-led work. It compares the decision record, work plan, and final runbook, then checks open tasks, wake-up mechanisms, and stalled sessions.

Rechecking project direction after context changes or major milestones, auditing active work, and identifying stalled or missing follow-up.

It helps detect forgotten tasks, false completion claims, inactive workers, and situations where two sessions are waiting on each other.

compass — Skill for Claude Code and Codex by 0fuk/OpenACCP

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.

Any agent

npx skills add 0fuk/OpenACCP --skill compass

Clone the repo

git clone --depth 1 https://github.com/0fuk/OpenACCP

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/0fuk/openaccp/compass/github.svg)](https://agentmods.dev/skills/0fuk/openaccp/compass)

Your own site

<a href="https://agentmods.dev/skills/0fuk/openaccp/compass"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/compass/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/0fuk/openaccp/compass"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/compass.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 126 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,841 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00126	$0.01841
Opus 5.5	$0.00050	$0.00736
Sonnet 5.5	$0.00025	$0.00368
Haiku 4.5	$0.00013	$0.00184

Measured 2d ago against content hash 734d25cc5322, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

compass 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.

field/skills/compass/SKILL.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/compass — 导航校准:对照目标 + 活性审计 + 跑偏自检

存在理由:主编排会冻在某时刻、叫醒机制清零、值守名存实亡;owner 不该靠人肉喊「看台账」把它拉回来。项目事实读项目根 PROJECT-PROFILE.md(台账/值守计划/心跳/协调目录),本 skill 零项目词。机械 hook 层见同目录 HOOKS.md。

模式

/compass=全检(六段);/compass 增量=§1-§3 只读上一份报告之后 mtime 有变的条目,§4/§5 全跑;/compass 轻检=只跑 §4+§5 一句话自答,无违规不落盘不刷屏。
档位纪律:compact/换代后/owner 点名=全检;大节点后=增量;周期循环=轻检。禁把该全检的场合降级。

1 对照目标(禁凭记忆)

真读三件套(决策台账/值守计划/终点 runbook,路径来自项目档案),引用目标态原文。compact 后记忆必稀释,「我记得目标是…」不合法。

2 对账(三栏,每条落点)

已完成:带证据路径;「做完了」不算。
在飞:执行体点名(子代理/外部会话/外部 CLI/候 owner)+叫醒机制。不点名=违规。
未动:带原因(候依赖/候 owner/明确降级)。

3 漏项扫描

①台账逐域扫已拍板未派工的条目;②文件信箱按 mtime 扫未消费件;③上一份报告的开放项逐条核销。漏项=当场入计划或候 owner 清单,「记下了」不是处置。

4 活性审计(核心,机械可判)

不变量:每个开放目标至少一个活的叫醒机制。

逐目标列叫醒机制;列不出=违规,当场架。
Monitor 判活只认通道实测(发一个真事件看它能否回到会话)(事件走完全程回到会话);标记文件/心跳新鲜不构成证据。存疑=无条件重架(幂等零成本)。
周期兜底=平台定时唤醒(ScheduleWakeup/loop),禁绑定会话进程的 while-loop;cron 在册≠触发(实证两次零触发),不作兜底。触发证据写独立文件(禁与 Monitor 心跳同文件,否则被覆写)。每次轻检第一动作=查上一 tick 的触发记录,缺=兜底已死,当场换机制并记违规。
速判式:存在开放目标 && 叫醒机制为零 → 无条件违规。
对端会话级硬检:凡执行体为其他会话的在飞项,活性只认其会话级副作用(进度行/回执/交回件 mtime);脱离进程写的心跳/TICK/标记文件只证进程活。会话级动作停更超 40 分钟→当场探活;探活 30 分钟无回执→备援(独立挑刺复核代理/临时复核)。
互等死锁判据:我在等 X 的件,且 X 最新会话级动作早于我的发令时刻 → 立即探活+推进不依赖它的活。
重启宽限:对端刚换代时心跳文件名可能沿用旧代,先看有无更新的 ACK/回执/SELFTEST,超一个完整周期未刷新才计停更。

5 自查有没有跑偏

当前每件事一句话答「对应台账哪条/runbook 哪段」,指不到=跑偏,停下归位或立项候裁。另查三病:

以计划结尾:turn 只许以「动作已发出/已收口/已挂叫醒机制」结束;
假的完成:宣称完成必须指向可验证产物;「测试全绿」≠决策完成度,自查对台账逐条;
等待合法性:等的是真依赖吗?不依赖它的产物是否已做完?
跳门检查(走五门发版范式的项目,轻检也跑):读现役五门台账(项目 CLAUDE.md 指向的最新工单集),逐卡看五门格——后一门有证据而前一门为空=跳门,当场上报并回填或退回;树上出现没有卡的改动=乱发明,停手立卡。 外部挑刺复核:完成度存疑/开放项对不上/连续两轮「全部正常」但 owner 有异议 → 派不带任何背景的子代理(sonnet,只给证据路径)独立重建三栏,不一致以子代理为准并上报。

6 落盘

全检报告落协调目录 <前缀>-COMPASS-导航-<日期HHMM>.md(目标态/完成/在飞/缺口/本轮修正),同步勾销值守计划。轻检只在有违规时落盘。

触发时机

owner 手动;compact/resume/换代后第一批动作;每个大节点后轻检;/keep-pushing 开窗时全检+架轻检循环+写值守开关文件;轻检循环 20-30 分钟一次。

4b · 子代理级活性硬检(20261010 事故后增补,轻检 / 增量 / 全检都跑)

判据:每个 running 子代理的 tasks\<agentId>.output mtime(只 ls,禁读内容)与跨会话执行体的回执 / 进行中件 mtime;距今 > 30 分钟 = 卡死,不是「在跑」。完成通知对卡死的执行体永远不会来,所以**「等完成通知」不算叫醒机制**;唯一合法的兜底是已挂的 ScheduleWakeup(≤ 30 分钟)。
速判式扩展:存在 running 子代理或未回执的跨会话令 && 无 ≤ 30 分钟 ScheduleWakeup 在册 → 无条件违规,当场挂表并记违规。
处置:卡死即 TaskStop → 按工作树态续派(见 /keep-pushing §7),harness 必含「禁 rm / scratchpad / 禁弹窗命令」三句;禁止等人点弹窗。
轻检输出一行:在飞 n:<id 短码>=<距上次动作分钟> …;兜底唤醒 <HH:MM>;任一 > 30 即违规落盘。
盲测收货:盲测交回收货前 blind-check exit=0(node C:/VyqoAI/Yirun/tools/blind-check.mjs <章程> <交回件> <证据目录>),否则违规落盘。

Read the full file on GitHub · 55 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

HOOKS.md 1.3 KB

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.

2d ago First seen · 55 lines · 126 tokens per session scan A 734d25cc5322

Subscribe to this mod's changes

compass is a skill published in the GitHub repository 0fuk/OpenACCP (2 stars, last pushed 2d ago), licensed MIT. It adds 126 tokens to every session and 1,841 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-10-14.

Building agents Multi-Agent Planning agentic-workflow ai-agents coordination handoff software-engineering

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…

Gentech-Labs/genTech-agent-kit · 71 tokens

agent-teams

Coordinates Claude agent teams via filesystem protocol. Use when orchestrating parallel agents with task dependencies. Do not use for single-agent tasks.

athola/claude-night-market · 30 tokens

orchestrate

This skill should be used when the user asks to 'orchestrate a task', 'break down work into parallel agents', 'coordinate subtasks', 'run agents in parallel', or mentions 'multi-agent'. Decomposes complex tasks into tracked subtasks, dispatches parallel subagents, and coordinates until completion.

varun29ankuS/shodh-memory · 67 tokens

compass