Linux Performance Analysis

What it does

Provides a systematic, safety-conscious skill for diagnosing Linux system performance issues. The skill covers baseline collection, targeted sweeps across CPU, memory, I/O, and network, and a drill-down process from system signals to specific processes. It includes rules for classifying kernel vs application bottlenecks and an explicit output contract for concise conclusions.

When to use it

Use when a host is slow, experiencing high load, swapping, timeouts, packet loss, or other performance degradations and you need a structured diagnostic to determine which subsystem is responsible and whether the cause is kernel- or application-level. Suitable for SREs, sysadmins, and support engineers.

What's included

• Scripts: none present in this fetch (has_scripts=false) — but the skill references multiple in-repo reference files in normal repositories; none were fetched here.
• References: none bundled in this fetch (has_references=false). The SKILL.md points to reference files (cpu, memory, io, network) which are expected in a fuller repo.
• Instructions: stepwise workflow (establish symptom, capture baseline, sweep areas, drill down, deliver conclusion) and escalation levels (L1 read-only to L3 intrusive sampling) with safety checks for production hosts.

Compatible agents

Best used by assistants with shell/run capabilities and read-only defaults; compatible with ops-focused agents (Copilot-style, Codex, Hermes) that can interpret command output and recommend safe next commands.