Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add lucifer1004/VeloQ --skill pytorch-profile-analysisgit clone --depth 1 https://github.com/lucifer1004/VeloQWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lucifer1004/veloq/pytorch-profile-analysis)<a href="https://agentmods.dev/skills/lucifer1004/veloq/pytorch-profile-analysis"><img src="https://agentmods.dev/badge/skills/lucifer1004/veloq/pytorch-profile-analysis/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/lucifer1004/veloq/pytorch-profile-analysis"><img src="https://agentmods.dev/badge/skills/lucifer1004/veloq/pytorch-profile-analysis.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00056 | $0.01389 |
| Opus 5 | $0.00028 | $0.00694 |
| Sonnet 5 | $0.00011 | $0.00278 |
| Haiku 4.5 | $0.00006 | $0.00139 |
Grade A, and why
pytorch-profile-analysis scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PyTorch Profile Analysis
Use veloq pytorch for PyTorch/Kineto Chrome traces:
veloq pytorch summary T
veloq pytorch search T --type kernel --name-regex 'nccl|gemm' --limit 20
veloq pytorch inspect T kernel:91
veloq pytorch correlate T kernel:91
veloq pytorch slices T --aggregate --group-by step
veloq pytorch collectives T
This skill requires the VeloQ CLI on PATH. If veloq is missing,
install it before analysis.
Tool Boundary
Use veloq pytorch verbs as the analysis interface. Do not query
<input>.veloq/pytorch/ sidecars, generated Parquet files, or raw
Kineto trace tables directly with DuckDB, PyArrow, pandas, or ad hoc
SQL unless the user explicitly asks for raw-trace exploration or you
are developing VeloQ itself.
veloq pytorch prep T only builds/checks sidecars. After prep, continue
with summary, search, inspect, stats, correlate, timeline,
slices, or collectives.
Inputs
- Explicit
veloq pytorchcommands accept one Chrome trace named.jsonor.json.gz. - Automatic source detection only claims
.pt.trace.jsonand.pt.trace.json.gz; explicitly selectpytorchfor other JSON filenames. - Directory inputs are not supported in PyTorch v0. Ask the user to choose one trace file if they point at a directory.
Row IDs
PyTorch row ids use <kind>:<stable_index>, where the stable index is
derived from the original traceEvents order after non-event flow markers
are skipped. Do not use Kineto Ev Idx as a stable key.
Use veloq pytorch schema <target> for the authoritative response field
inventory; do not infer the public contract from raw Kineto fields.
Common prefixes:
| Type | Row id prefix |
|---|---|
| CPU op | cpu_op:N |
| Annotation | annotation:N |
| Step | step:N |
| Runtime | runtime:N |
| Driver | driver:N |
| Kernel | kernel:N |
| Memcpy | memcpy:N |
| Memset | memset:N |
| Memory | memory:N |
| Python | python:N |
| Comm | comm:N |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 157 lines · 56 tokens per session scan A 0867b00f768e
pytorch-profile-analysis is a skill published in the GitHub repository lucifer1004/VeloQ (120 stars, last pushed 20d ago), licensed MIT. It adds 56 tokens to every session and 1,389 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
modal-serverless-gpu
Run approved CPU or GPU work through OpenScience computejob on the user's configured Modal account. Use for isolated scientific scripts, dependency provisioning, durable outputs, logs, status, cancellation, and recovery. Never invoke the Modal SDK or CLI directly.
cuopt-developer
Modify, build, test, debug, and contribute to NVIDIA cuOpt (C++/CUDA, Python, server, CI). Use for solver internals, PRs, DCO, and code conventions.
cuopt-user-rules
Base rules for end users calling NVIDIA cuOpt (routing/LP/MILP/QP/install/server). Not for cuOpt internals — use cuopt-developer for those.
ncu-cuda-profiling
Automated NCU (Nsight Compute) profiling workflow with full metrics collection and persistent storage.
cuda-kernel-optimizer
Use when optimizing or diagnosing GPU workload performance with CUDA, CUTLASS, Triton, PyTorch, vLLM or TensorRT-LLM; interpreting NCU, Nsys or PyTorch Profiler evidence; or preparing the workload and correctness checks for such optimization. General coding, skill maintenance and conceptual GPU questions do not…
cuopt-numerical-optimization-formulation
LP, MILP, QP — concepts, problem-text parsing, and formulation patterns (parameters, constraints, decisions, objective). Concepts only; no API.