Download for MacApple silicon, macOS 14+
Skill

ns-cpu-spike-analysis

@nodesource Updated 2026-09-16

Diagnoses high CPU or slow execution in a Node.js app using supplied evidence or N|Solid CPU metrics/profiles. Use when the user reports CPU spike, high CPU percent, slow endpoint/function, event loop blocking, flamegraph/profile analysis, hot loop, or asks why an app is slow. Prefer existing asset IDs, summaries, trace data, or local profiles before capturing a new profile.

Install

With your agent

Install OMGSkills skill "nodesource/nsolid-skills-plugin:skills/ns-cpu-spike-analysis" into this agent's normal skills directory: call the OMGSkills MCP get_install_plan tool for your agent and run the script it returns (it installs only the pinned commit and verifies SKILL.md), and stop if MCP or the pin is unavailable. Do not overwrite or change existing skills without asking me.

Manual command

git clone https://github.com/nodesource/nsolid-skills-plugin /tmp/nsolid-skills-plugin && ln -s /tmp/nsolid-skills-plugin/skills/ns-cpu-spike-analysis ~/.claude/skills/ns-cpu-spike-analysis

From README

Use Provided Evidence First Treat the prompt content as primary evidence. Parse any provided app name, agent ID, hostname, time window, asset ID, local file path, CPU summary, flamegraph excerpt, hot function list, or stack trace before calling tools. If the prompt already includes an asset ID, local .cpuprofile path, or structured CPU summary, analyze that evidence first instead of starting with information-dashboard or metrics-historic. If the prompt is a telemetry alert such as "CPU spike 121.1% in app X", use that app name as authoritative scope and immediately run the live workflow: identify the hottest connected agent inside that app, capture a 30-second CPU profile, summarize it, download it, and fetch runtime code when possible. In telemetry-alert mode, do not stop after rediscovering the same spike and do not ask for capture approval unless the prompt explicitly says read-only, offline, or no-capture.