菜单
Collect system and cognitive health metrics. Returns a structured health report with status classifications (ok/warning/critical) for hardware, OS, and cognitive subsystems.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/bdambrosio/Cognitive_workbench --skill check-health命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Notices when someone comes near you or moves away in the shared world. Fires on the transition only, and carries the current situation with it.
Look around the shared world — the ground you are standing on, and who else is present, how far away they are and whether they are facing you. Your primary perception when you are embodied in the world.
Leave a labelled marker in the shared world, at your feet or at a coordinate. Markers persist and are visible to anyone who can see that spot — a way to point at a place without saying a coordinate out loud.
基于 SOC 职业分类
正在显示 SKILL.md
| name | check-health |
| type | python |
| description | Collect system and cognitive health metrics. Returns a structured health report with status classifications (ok/warning/critical) for hardware, OS, and cognitive subsystems. |
| schema_hint | {"out":"$variable (optional)"} |
Homeostasis monitor. Collects hardware metrics (GPU, CPU, RAM, disk, network, processes) via system-health.sh and cognitive metrics (memory integrity, tool effectiveness, world model health, planning performance) from internal state. Returns a structured health report.
No required input. Runs all checks automatically.
Success (status: "success"):
value: Human-readable summary string — lists concerns if any, otherwise "All systems nominal."$health) contains a JSON object with these top-level keys:
overall_status: "ok" | "warning" | "critical"concerns: list of human-readable concern strings (empty if all ok)system: hardware/OS metrics (cpu, gpu, ram, disk, network, processes)cognitive: cognitive metrics (memory, planning, tool_model, world_model, user_interaction)health = executor.get_json("$health")
overall = health["overall_status"] # "ok", "warning", or "critical"
concerns = health["concerns"] # e.g. ["Plan completion rate low: 45% (last 20 plans)"]
Do NOT use health.get("extra", {}).get("overall_status") — overall_status and concerns
are at the top level of the Note content, not nested under extra.
src/scripts/system-health.sh synchronously to collect OS/hardware metricsScheduled health check (planner decides what to report):
{"type":"check-health","out":"$health"}
The planner reads $health via executor.get_json("$health") and decides whether to
say or ask the user based on overall_status and concerns.