AgentSkillsCN

system-triage

针对系统与网络问题的标准化诊断流程。

SKILL.md
--- frontmatter
name: system-triage
description: Standardized diagnotic workflow for system and network issues.
role_required: true
tools:
  - run_command
  - send_command_input

System Triage Skill

Prerequisites

  • •Active Role: Must be in a Role Context (e.g., //office @it).
  • •Issue Context: Must be working on a specific issue (Jira/Gitea).

Workflow

1. Acknowledge

  • •Action: Comment on the issue acknowledging receipt.
  • •Message: "Beginning investigation."

2. Define Problem

  • •Check: Are timestamps, source IPs, and error messages present?
  • •Action: If ambiguous, comment asking for clarification. STOP.

3. Live Investigation (Scout)

Execute sequentially. Log every output.

3a. Network Socket Investigation

  • •Tool: ss
  • •Command: ss -lupn | grep '<port>'
  • •Goal: Identify listening processes.

3b. Process Investigation

  • •Tool: ps
  • •Command: ps aux | grep -E -i '<keyword>'
  • •Goal: Identify running processes by name.

3c. Service Status Check

  • •Tool: systemctl
  • •Command: systemctl status <service>
  • •Goal: detailed status, config paths, logs.

4. Formulate Hypothesis

  • •Action: Analyze data from Step 3.
  • •Output: meaningful summary comment with evidence-based hypothesis.

5. Propose Action Plan

  • •Option A (Change Required): Propose specific system modification. Trigger infra-change-request.
  • •Option B (More Info): Outline specific next steps for deeper investigation.