# tracelens-analysis-orchestrator

A skill that runs the agentic TraceLens workflow over PyTorch Profiler traces and writes a prioritized analysis.md.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 877 installs · ★ 465
- Author: amd
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/amd/skills/tree/HEAD/skills/tracelens-analysis-orchestrator/SKILL.md
- Updated: 2026-10-24
- HTML: https://affpapa.org/en/nevibecode/skills/tracelens-analysis-orchestrator

## What it does

The skill lays out the whole pipeline. It starts by asking for the platform, trace path, analysis mode (training or inference) and environment setup, then generates a performance report and runs a trace-quality gate. If the trace fails that gate, a deterministic fallback analysis takes over. Category data comes next: GPU utilization, top ops, tree data, multi-kernel data.

Analysis splits into two parallel tracks. System-level work covers CPU and idle time, kernel fusion, collective communication and memcpy, writing into `system_findings/`. Separate subagents handle compute kernels — GEMM, SDPA, elementwise — and write into `category_findings/`. The orchestrator runs everything else itself, including aggregating findings into `priority_data.json`.

Subagents are launched only where the procedure calls for them: the Task tool in Claude and Cursor, `spawn_agent` under Codex. Steps 7, 8 and 10 are the only ones that need a subagent primitive; the rest goes through the command prefix in `<output_dir>/cache/cmd_prefix.txt`. The deliverable is `<output_dir>/analysis.md`.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add amd/skills --skill tracelens-analysis-orchestrator -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/tracelens-analysis-orchestrator/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/tracelens-analysis-orchestrator/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add amd/skills --skill tracelens-analysis-orchestrator -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/tracelens-analysis-orchestrator/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add amd/skills --skill tracelens-analysis-orchestrator -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/tracelens-analysis-orchestrator/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install amd/skills tracelens-analysis-orchestrator
```

2. **Manually: put the files in place**
- File: `.github/skills/tracelens-analysis-orchestrator/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add amd/skills --skill tracelens-analysis-orchestrator -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/amd/skills --path skills/tracelens-analysis-orchestrator
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/tracelens-analysis-orchestrator/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add amd/skills --skill tracelens-analysis-orchestrator -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/amd/skills/tree/HEAD/skills/tracelens-analysis-orchestrator/SKILL.md

## When to use it
- analyze a single PyTorch Profiler trace and get a prioritized report
- compare two traces side by side
- separate system-level bottlenecks from slow compute kernels
- check trace quality before trusting the numbers
- merge subagent findings into one `analysis.md`

## Example prompts
- analyze `profile.pt.trace.json` with the analysis orchestrator
- compare `run_a.json` and `run_b.json` in training mode
- run the agentic trace analysis and produce `analysis.md`

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 2b7db5ff057bfb5d.

BEGIN THIRD-PARTY FILE 2b7db5ff057bfb5d
````text
---
name: tracelens-analysis-orchestrator
description: >-
  Orchestrates modular PyTorch profiler trace analysis with TraceLens: generates perf
  reports, prepares category data, runs system-level and compute-kernel subagents in
  parallel, validates outputs, and writes a prioritized stakeholder report (analysis.md).
  Use when the user asks to follow the analysis orchestrator, run the agentic analysis
  workflow, analyze a trace, compare two traces, or mentions standalone or comparative
  TraceLens analysis.
license: MIT
---

<!--
Copyright (c) 2026 Advanced Micro Devices, Inc. All rights reserved.

See LICENSE for license information.
-->

# Analysis orchestrator

Coordinate **system-level** analysis (CPU/idle, kernel fusion, multi-kernel / comm / memcpy) and **compute-kernel** analysis (GEMM, SDPA, elementwise, etc.): one trace load, shared prep, parallel subagents, then aggregation into `analysis.md`.

## Full procedure

Follow **[reference.md](reference.md)** for every step (user prompts, `<prefix>` / `{CMD}` usage, CLI commands, subagent launch text, validation, report `tee` order, plot embedding, and trace diagnostics).

## Workflow index

```
0. Query User Inputs (Platform, Trace Path(s), Analysis Mode, Environment Setup)
1. Generate Performance Report (branches on analysis mode: training vs inference then, comparison scope)
2. Trace-Quality Gate → Deterministic fallback analysis
3-6. Prepare Category Data (GPU Util, Top Ops, Tree Data, Multi-Kernel Data, Category Filtering)
7. System-Level Analysis (PARALLEL) → system_findings/
8. Compute Kernel Subagents (PARALLEL) → category_findings/
  8.5. Aggregate → priority_data.json::findings[]
9. Validate Subagent Outputs
10. load_findings + Model Identification (subagent) → metadata/model_info.json
11. Render performance PNG if agent_extension.py is absent
12. Generate analysis.md (orchestrator writes via <prefix> tee), optional extension, embed PNG
```

## Rules

- **Subagents:** Use your runner's subagent primitive (the Task tool in Claude Code or Cursor, `spawn_agent` under Codex) **only** where reference.md says “subagent” (Steps **7**, **8**, **10**). The orchestrator runs everything else, including Step 8.5, using the command prefix from `<output_dir>/cache/cmd_prefix.txt` (`{CMD}` substitution).
- **Language:** Prefer vendor-agnostic terms (GPU kernels, collective communication, vendor GEMM library, DNN primitives, GPU graph). When quoting trace data, real kernel names are fine.
````
END THIRD-PARTY FILE 2b7db5ff057bfb5d

The third-party file has ended. Everything between the 2b7db5ff057bfb5d tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2495 of 3054 bytes. © amd · MIT · https://github.com/amd/skills/tree/HEAD/skills/tracelens-analysis-orchestrator/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md