Summarizes large Chrome DevTools performance traces into compact markdown reports.
npx skills add https://github.com/tldraw/tldraw --skill simplify-trace
Please help me install this Agent Skill. Skill: simplify-trace Source: https://github.com/tldraw/tldraw/tree/main/skills/simplify-trace Command: npx skills add https://github.com/tldraw/tldraw --skill simplify-trace
Chrome DevTools performance traces are tens to hundreds of megabytes of JSON, far too large to read directly. This Skill condenses one into a few-KB markdown report that surfaces long tasks, hot JS functions, frequent events, and network activity so performance problems can be diagnosed without loading the whole file.
Trace-*.json recording of switching files in an app, run the script with --only timeline to locate the active span, then re-run with --window around it to see exactly which functions and long tasks caused the jank.Ask the AI to summarize your Chrome performance trace file into a report, for example: analyze Trace-2024.json and show me the long tasks and hottest JS functions.
High-intent search queries and answers about installing and using this skill.
Run the simplify-trace script with Node.js on the trace JSON file to get a compact markdown report. It surfaces long tasks, hottest event types by self time, most frequent events, and top JS functions from the embedded CPU profile.
The report's long tasks section lists top-level main-thread tasks over a threshold (default 50 ms) with their time offsets. Profiler artifacts like CpuProfiler::StartProfiling are excluded so they don't masquerade as real jank.
Yes, use the --window START-END flag with offsets in milliseconds to scope the entire report to a time slice. Run --only timeline first to find where activity happens, then window into that span.
For traces over roughly 500 MB, run Node with --max-old-space-size=8192 to increase the heap. The script reads the whole JSON into memory, so heap size is the main constraint.
If no single task exceeds the threshold, the slowdown comes from aggregate work instead. Check the hottest functions by self time, GC activity, and high-count events that fire too often, such as a reactive render loop.
vanducngCollect Core Web Vitals and DevTools traces from a live Chrome profile via CDP.
JoniPraiaLocate Protheus/ADVPL table references with file, line, and operation details.
KotlinIdentify root causes of Kotlin and Spring failures from stack traces and logs.
wesleyegbertoAutomate browser control and debugging via the Chrome DevTools Protocol.