Borrowing it
Nothing to install: this file belongs to NVIDIA/TensorRT-LLM. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/NVIDIA/TensorRT-LLM/main/.claude/skills/perf-regression-cookbook/SKILL.mdgit clone --depth 1 https://github.com/NVIDIA/TensorRT-LLMWrote 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/nvidia/tensorrt-llm/perf-regression-cookbook)<a href="https://agentmods.dev/skills/nvidia/tensorrt-llm/perf-regression-cookbook"><img src="https://agentmods.dev/badge/skills/nvidia/tensorrt-llm/perf-regression-cookbook/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/nvidia/tensorrt-llm/perf-regression-cookbook"><img src="https://agentmods.dev/badge/skills/nvidia/tensorrt-llm/perf-regression-cookbook.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00110 | $0.06024 |
| Opus 5.5 | $0.00044 | $0.02410 |
| Sonnet 5.5 | $0.00022 | $0.01205 |
| Haiku 4.5 | $0.00011 | $0.00602 |
Grade A, and why
perf-regression-cookbook 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 17d 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.
The source is not reproduced here
A licence we could not identify
The repository carries a LICENSE file, but it is custom or dual enough that GitHub cannot name it and neither can this catalogue. Unknown terms are not permission, so the body is not copied here. Read the licence at the source and decide for yourself.
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- data/aliases.yaml 5.6 KB
- data/modules.yaml 7.7 KB
- data/patterns.yaml 32 KB
- data/pending.yaml 16 KB
- data/removed.yaml 11 KB
- data/tags.yaml 6.9 KB
- references/attention-fmha/dsa-indexer-host-overhead.md 4.4 KB
- references/attention-fmha/dsa-indexer-tp-duplicated-prefill.md 3.6 KB
- references/attention-fmha/dsmla-sparse-ctx-2cta-fallback.md 6.6 KB
- references/attention-fmha/fmha-dispatcher-sm-count-query.md 3.4 KB
- references/attention-fmha/index.md 3.9 KB
- references/attention-fmha/trtllmgen-mla-decode-unwaived-slower.md 5.7 KB
- references/autotuner/index.md 1.3 KB
- references/autotuner/nested-nvfp4-gemm-autotune-dispatch.md 4.1 KB
- references/case-template.md 8.0 KB
- references/communication/allreduce-host-overhead-small-model-tp.md 8.7 KB
- references/communication/allreduce-strategy-lut.md 4.4 KB
- references/communication/index.md 3.4 KB
- references/communication/nemotronh-mnnvl-nccl-fallback.md 3.1 KB
- references/cuda-graph-and-compile/cudagraph-default-batch-sizes.md 3.7 KB
- references/cuda-graph-and-compile/index.md 3.4 KB
- references/cuda-graph-and-compile/mrope-graph-gate-text-only-eager.md 4.8 KB
- references/cuda-graph-and-compile/piecewise-attention-torch-compile-host-overhead.md 4.4 KB
- references/cuda-graph-and-compile/piecewise-cudagraph-capture-coverage.md 4.5 KB
- references/cuda-graph-and-compile/piecewise-cudagraph-far-ceiling-append.md 5.7 KB
- references/disagg-orchestrator/disagg-fill-throttle-slow-start.md 3.4 KB
- references/disagg-orchestrator/disagg-server-gc-pauses.md 4.5 KB
- references/disagg-orchestrator/index.md 1.9 KB
- references/gemm-and-quantization/deepgemm-pdl-import-time-cuda-context.md 7.6 KB
- references/gemm-and-quantization/fp8-pertensor-quant-slower-than-te.md 3.8 KB
- references/gemm-and-quantization/index.md 2.2 KB
- references/jit-and-warmup/index.md 4.1 KB
- references/jit-and-warmup/maybe-compile-recompiles-every-call.md 4.5 KB
- references/jit-and-warmup/mla-chunked-prefill-maybe-compiled-cat-warmup.md 5.0 KB
- references/jit-and-warmup/model-engine-warmup-orchestration-gaps.md 4.5 KB
- references/jit-and-warmup/trtllmgen-fmha-densify-grid.md 6.5 KB
- references/jit-and-warmup/trtllmgen-fmha-jit-warmup.md 5.1 KB
- references/jit-and-warmup/warmup-token-cap-revert.md 5.3 KB
- references/kernel-fusion/flinear-bias-copy-sm100.md 3.0 KB
- references/kernel-fusion/index.md 2.4 KB
- references/kernel-fusion/qwen35-fused-ops-disabled-under-torch-compile.md 4.3 KB
- references/kv-cache-manager/dsa-kcache-dtype-size-estimate.md 3.3 KB
- references/kv-cache-manager/index.md 3.6 KB
- references/kv-cache-manager/mamba-hybrid-recurrent-state-d2h-syncs.md 5.7 KB
- references/kv-cache-manager/max-utilization-reuse-token-budget.md 4.9 KB
- references/kv-cache-manager/qwen35-ssm-cache-fp32-fallback.md 4.7 KB
- references/kv-cache-transceiver/disagg-gen-retokenize-and-memcpy.md 5.1 KB
- references/kv-cache-transceiver/disagg-quorum-vote-world-scope.md 3.8 KB
- references/kv-cache-transceiver/index.md 2.0 KB
- references/measurement-and-test/index.md 5.0 KB
- references/measurement-and-test/ipc-hmac-key-via-fd-breaks-bench.md 6.9 KB
- references/measurement-and-test/multi-gpu-bench-warmup-too-few.md 3.4 KB
- references/measurement-and-test/nixl-ctx-only-gap-not-reproduced-home-mount.md 3.4 KB
- references/measurement-and-test/perf-sanity-sampler-options-applied-to-all-cases.md 3.5 KB
- references/measurement-and-test/perf-test-gpt-oss-20b-moe-backend-pin.md 4.7 KB
- references/measurement-and-test/perf-test-wrong-moe-backend.md 4.1 KB
- references/model-definition/index.md 2.7 KB
- references/model-definition/ltx2-bf16-lora-restore.md 4.3 KB
- references/model-definition/mamba2-flashinfer-head-group-ratio-gate.md 6.1 KB
- references/model-definition/vlm-vision-tower-loaded-for-text-only-bench.md 7.2 KB
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.
- 17d ago First seen · 351 lines · 110 tokens per session scan A e18d82de8176
perf-regression-cookbook is a skill published in the GitHub repository NVIDIA/TensorRT-LLM (15,021 stars, last pushed today), with no licence file. It adds 110 tokens to every session and 6,024 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-10-20.
perf-regression-cookbook compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| perf-regression-cookbookNVIDIA/TensorRT-LLM | Skill | 15k | today | none | A | 110 | Claude Code |
| sglang-diffusion-benchmark-profilesgl-project/sglang | Skill | 39k | today | Apache-2.0 | A | 26 | Claude Code, Codex |
| sglang-diffusion-performancesgl-project/sglang | Skill | 39k | today | Apache-2.0 | A | 29 | Claude Code, Codex |
| sglang-diffusion-add-modelsgl-project/sglang | Skill | 39k | today | Apache-2.0 | A | 24 | Claude Code, Codex |
| sglang-diffusion-cache-ditsgl-project/sglang | Skill | 39k | today | Apache-2.0 | A | 161 | Claude Code |
| sglang-diffusion-modelopt-quantsgl-project/sglang | Skill | 39k | today | Apache-2.0 | A | 49 | Claude Code, Codex |
| doctortinyfish-io/tinyfish-cookbook | Skill | 2.3k | 3d ago | MIT | A | 72 | Claude Code |