Profile a hot path (feature + backend) using the appropriate profiler (perf / ncu / Vtune / rocprof), produce flamegraph + top-N hot functions, and suggest concrete optimizations. Delegates to the perf-profiler agent.
Scanned 9/12/2026
Install to Claude Code
npx -y skills add VMAFx/vmafx --skill profile-hotpath --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Profile Hotpath?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vmafx-profile-hotpath)More formats (shields.io, HTML) on the badges page.
---
name: profile-hotpath
description: Profile a hot path (feature + backend) using the appropriate profiler (perf / ncu / Vtune / rocprof), produce flamegraph + top-N hot functions, and suggest concrete optimizations. Delegates to the perf-profiler agent.
---
<!-- markdownlint-disable MD013 -->
# /profile-hotpath
## Invocation
```text
/profile-hotpath <backend> <feature> [--resolution=576|640|720|1080|4k] [--frames=120]
```
## Behavior
Delegates the profiling + interpretation work to the `perf-profiler` agent (see
`.claude/agents/perf-profiler.md`), which:
1. Builds (if needed) with `--config=relwithdebinfo` for symbolicated profiles.
2. Selects the profiler for the backend.
3. Runs `build/tools/vmaf_bench` with fixed seed + frame count.
4. Produces a flamegraph under `build/profiles/<date>/flame.svg`.
5. Emits a top-10 hot-function list + recommendations.
## Notes
- The `perf-profiler` agent owns the actual interpretation — this skill just sets up
the environment and invokes it.
- If the profiler isn't installed (`ncu` missing for CUDA, `vtune` for SYCL), the skill
degrades to `perf record` as a fallback, with a clear note in the report.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.