Auto Monitor - System Monitoring

主动监控系统状态。定期检查服务器健康，主动汇报，无需等待指令。

Security Scan

The skill's stated purpose (automatic system monitoring) matches its instructions, but the runtime guidance is vague about what checks/repairs will be executed and grants the agent broad, potentially destructive discretion without declaring required privileges or safeguards.

Purpose & Capability

Name/description and SKILL.md align: this is a proactive system-monitoring skill. However, it does not state what runtime permissions, binaries, or account context it expects to use to inspect or modify the system — a legitimate monitor normally documents required commands/privileges or an install mechanism for agents to run safely.

Instruction Scope

Instructions tell the agent to perform periodic checks (disk, memory, CPU, network, processes) and to '自动修复' (auto-fix) issues. The SKILL.md gives no concrete commands, limits, or definition of 'auto-fix', giving the agent broad discretion to read system state and to modify/restart services or files. That vagueness enables unsafe or overly-broad actions.

Install Mechanism

No install spec and no code files — lowest installation risk. The skill is instruction-only so nothing is written to disk by an installer.

Credentials

The skill declares no required env vars but implies access to system-level information and possible modifications. It does not explain needed credentials or account privileges (e.g., root vs unprivileged), nor whether it will read config files or secrets. This mismatch between implied capabilities and declared requirements is a proportionality concern.

Persistence & Privilege

always is false, but disable-model-invocation is false (default), and the skill's purpose is autonomous, periodic monitoring with automatic remediation. Combined, that allows the agent to repeatedly invoke itself and take remediation actions without clear limits or user confirmation — a higher-risk operational profile unless restricted.

What to consider before installing

This skill could act on your system automatically and the SKILL.md is vague about what 'auto-fix' does or what permissions it needs. Before installing or enabling autonomous invocation: 1) Require the author to provide exact commands/actions the agent will run for each check and for each automatic remediation (e.g., 'restart service X', 'delete temp files older than Y', 'clear cache'); 2) Require a clear list of privileges/accounts needed (avoid root if possible); 3) Demand safeguards: dry-run mode, confirmation before destructive actions, detailed logs, and notification channels; 4) If you must test, run it in a sandbox or on a non-production system under a restricted user account; 5) Prefer a version that includes explicit code or an install script you can review rather than an instruction-only skill with open-ended authority. If the author cannot clarify these points, treat the skill as risky and avoid enabling autonomous execution.

Auto Monitor

主动监控系统，主动发现问题。

能力轮廓

• 输入：系统状态
• 输出：健康报告 + 异常告警
• 核心：主动检查 → 及时汇报

工作流

1. 定期检查（每 N 分钟）
2. 检查项：
   - 磁盘使用率
   - 内存使用率
   - CPU 负载
   - 网络状态
   - 进程状态
3. 异常判断
   - 超过阈值？→ 主动告警
4. 汇报
   - 正常：简洁汇报
   - 异常：详细说明 + 建议

告警阈值

• 磁盘 > 80%
• 内存 > 85%
• CPU Load > 核数 × 0.8

主动性

• 不等用户问"服务器怎么样"
• 发现问题主动说
• 可以自动修复的立刻处理