New: 97,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Torch Tensorrt

by VectorSpaceLab in VectorSpaceLab/AREX-Skill

A skill your agent uses for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging…

BSD-3-ClauseAuto-check passedAI & LLM Engineering

Install Torch Tensorrt

skills CLI
$ npx skills add VectorSpaceLab/AREX-Skill --skill torch-tensorrt -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install VectorSpaceLab/AREX-Skill torch-tensorrt --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/VectorSpaceLab/AREX-Skill.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/repositories/repo-skills/torch-tensorrt .claude/skills/torch-tensorrt && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
torch-tensorrt
GitHub stars
328
Token cost
~1.5k tokens
SKILL.md length
619 words
Files
8 (incl. scripts, references)
Skills in repo
159
Repo updated
First seen
Licence
BSD-3-Clause

At a glance

A skill your agent uses for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging…

  • Works in 6 steps: Confirm the user has, or is willing to… → If the package is missing, start with a… → Run the bundled environment probe when… → …
  • Torch-TensorRT tasks: compiling PyTorch models with TensorRT
  • SKILL.md covers When to use this skill, First checks, Route by task and Repository-specific operating…, plus 1 more section
  • Runs Python scripts from its folder; calls python

What it does

Torch Tensorrt is an agent skill from VectorSpaceLab/AREX-Skill. Use this skill for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging unsupported ops, and maintaining source builds.

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 9 other files, including scripts and reference files (for example `references/api-surface-map.md`, `references/installation-and-features.md` and `references/repo-provenance.md`).

It sits in AI & LLM Engineering, covering LLM inference and serving and Deep learning. It works with NVIDIA AI Platform, PyTorch, C++ and CUDA. The repository describes itself as: A Skill Library for Automated Machine Learning. The licence is BSD-3-Clause.

When your agent uses it

  • Torch-TensorRT tasks: compiling PyTorch models with TensorRT
  • Dynamic-shape/export workflows
  • Runtime optimization
  • Triton/C++/distributed deployment

Example prompts

  • “/torch-tensorrt”

Requirements

  • Python 3

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. Confirm the user has, or is willing to prepare, an NVIDIA GPU runtime with compatible CUDA, PyTorch, TensorRT or TensorRT-RTX, and…
  2. If the package is missing, start with a matching public install command and then narrow the environment from there
  3. Run the bundled environment probe when install state is uncertain
  4. Read references/installation-and-features.md before giving install advice, choosing standard TensorRT versus TensorRT-RTX, diagnosing…
  5. Read references/api-surface-map.md when the user names an API but it is unclear which sub-skill owns it.
  6. Read references/troubleshooting.md when the first visible symptom is an import, CUDA, TensorRT wheel, serialization, quantization, or…

What it can do on your machine

Read from SKILL.md and the folder at commit ac3fe1a. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Torch Tensorrt loads about 1.5k tokens when it runs, and up to ~6.9k if it reads all its reference files. Until then it costs about 61 tokens; SKILL.md has 619 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~61
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~6.9k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from VectorSpaceLab/AREX-Skill at commit ac3fe1a, republished under its BSD-3-Clause licence (© VectorSpaceLab). 619 words, ~1,502 tokens.

Download SKILL.mdSave it as .claude/skills/torch-tensorrt/SKILL.md (or your agent's skills folder). This skill also uses 7 other files; get the full folder from GitHub.
name
torch-tensorrt
description
Use this skill for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging unsupported ops, and maintaining source builds.
metadata.disco-role
operating
disable-model-invocation
true
license
BSD 3-Clause

Torch-TensorRT

Torch-TensorRT accelerates PyTorch inference on NVIDIA GPUs by compiling supported graph regions into TensorRT engines while keeping PyTorch integration for the rest of the workflow. Use this root as a router first; load sub-skills for detailed procedures.

When to use this skill

Use this skill when the request mentions any of these signals:

  • Torch-TensorRT, torch_tensorrt, torch-tensorrt, torch-tensorrt-rtx, TensorRT-RTX, backend="torch_tensorrt", ir="dynamo", or torchtrtrun.
  • PyTorch-to-TensorRT inference compilation, torch.export plus TensorRT, TensorRT engine serialization, .ep, .ts, .pt2, .engine, or .pte artifacts.
  • Dynamic TensorRT input shapes, optimization profiles, Input(...), Device(...), precision settings, unsupported-op fallback, dryrun, custom converters, TensorRT plugins, or QDP kernels.
  • CUDA Graphs, engine caches, runtime settings, mutable Torch-TensorRT modules, refit, weight streaming, Triton serving, C++ runtime, ExecuTorch, DLA/Jetson, Windows cross-compile, or distributed inference.

Do not use this skill for generic PyTorch training, generic TensorRT C++ applications with no PyTorch/Torch-TensorRT layer, or unrelated model-serving frameworks unless Torch-TensorRT artifacts are part of the task.

First checks

  1. Confirm the user has, or is willing to prepare, an NVIDIA GPU runtime with compatible CUDA, PyTorch, TensorRT or TensorRT-RTX, and torch_tensorrt installed.

  2. If the package is missing, start with a matching public install command and then narrow the environment from there:

    bash
    python -m pip install torch torch-tensorrt tensorrt
# or, for the RTX variant
python -m pip install torch torch-tensorrt-rtx

    Adjust CUDA/PyTorch versions to the user's platform before promising success.

  3. Run the bundled environment probe when install state is uncertain:

    bash
    python scripts/check_torch_tensorrt_env.py --no-cuda-smoke

    The script is safe: it imports packages, prints versions/features, and only allocates a tiny CUDA tensor when CUDA is visible.

  4. Read references/installation-and-features.md before giving install advice, choosing standard TensorRT versus TensorRT-RTX, diagnosing optional features, or interpreting ENABLED_FEATURES.

  5. Read references/api-surface-map.md when the user names an API but it is unclear which sub-skill owns it.

  6. Read references/troubleshooting.md when the first visible symptom is an import, CUDA, TensorRT wheel, serialization, quantization, or unsupported-op error.

Route by task

User goalLoad this sub-skill
Compile a model through torch.compile, torch_tensorrt.compile, torch.export, dynamic shapes, precision/settings, save/load, raw engine extraction, or Windows cross-compile API choicessub-skills/compilation-and-export/SKILL.md
Optimize a compiled module at runtime: CUDA Graphs, output allocator, preallocated outputs, weight streaming, engine/timing/runtime caches, TensorRT-RTX runtime settings, mutable modules, refit, benchmarking, memory triagesub-skills/runtime-optimization/SKILL.md
Deploy compiled artifacts to Python, C++, AOTInductor, Triton, ExecuTorch, DLA/Jetson, Windows/ARM64, TensorRT-RTX, or distributed inference with torchtrtrunsub-skills/deployment-and-distributed/SKILL.md
Debug or extend compiler behavior: dryrun, Debugger, capture/replay, unsupported ops, converter/lowering/plugin authoring, QDP kernels, ModelOpt/quantization warnings, issue-quality reprossub-skills/extensibility-and-debugging/SKILL.md
Build or test Torch-TensorRT from source, choose package variants, inspect CI/test lanes, or maintain repository codesub-skills/build-and-maintenance/SKILL.md
Show full SKILL.md (219 more words)Show less

Repository-specific operating rules

  • Prefer the Dynamo path for new Python workflows: torch.compile(..., backend="torch_tensorrt"), torch_tensorrt.compile(..., ir="dynamo"), or torch.export.export(...) followed by torch_tensorrt.dynamo.compile(...).
  • Treat the legacy TorchScript/C++ APIs as real but specialized. Route them through deployment or build/maintenance when the task explicitly needs .ts, C++/libtorch, DLA, or source builds.
  • Always state backend prerequisites and feature gates before promising execution. Standard TensorRT, TensorRT-RTX, QDP kernels, ModelOpt quantization, distributed/NCCL, ExecuTorch, and C++ runtime features have different package and platform requirements.
  • Do not instruct the user to run examples or scripts from an original source checkout. Use the bundled references and scripts in this skill, or write task-local code for the user.
  • For performance claims, require user-side measurement with warmups, CUDA synchronization/events, representative dynamic shapes, and a PyTorch baseline. Do not infer speedups from compile success alone.
  • For unsupported operators, first decide whether fallback is acceptable. Use dryrun, torch_executed_ops, min_block_size, or model rewrites before recommending custom converter/plugin work.

Evidence and refresh

  • Source snapshot and evidence map: references/repo-provenance.md.
  • Router metadata for managed imports: references/repo-routing-metadata.json.
  • This skill was generated from source and documentation evidence plus a partial installed-package inspection. The inspection proved imports and a tiny TensorRT-RTX Dynamo compile, but did not prove all standard TensorRT, C++ runtime, serialization, distributed, QDP, or ModelOpt paths. Preserve those limits in downstream advice unless the user's current environment verifies them.

© VectorSpaceLab, BSD-3-Clause. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 7 other files (scripts, references) in skills/repositories/repo-skills/torch-tensorrt of VectorSpaceLab/AREX-Skill.

  • SKILL.md
  • references/api-surface-map.md
  • references/installation-and-features.md
  • references/repo-provenance.md
  • references/repo-routing-metadata.json
  • references/troubleshooting.md
  • scripts/check_torch_tensorrt_env.py
  • sub-skills

Open the folder on GitHubat commit ac3fe1a

Compare with similar skills

Torch Tensorrt next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Torch Tensorrt compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Torch Tensorrt this skillVectorSpaceLab/AREX-Skill328—~1.5kAutomated safety check: PassBSD-3-Clause
Graphsignalgraphsignal/graphsignal257—~6.2kAutomated safety check: PassApache-2.0
Quark Env Preflightamd/Quark181—~1.4kAutomated safety check: PassMIT
Spark Environment Setupwshobson/agents40k—~2kAutomated safety check: PassMIT
Ako4allTongmingLAIC/AKO4ALL369—~4kAutomated safety check: PassMIT
Paddle Op DevPaddlePaddle/Paddle24k—~1.3kAutomated safety check: PassApache-2.0

Similar skills

  • Graphsignal

    graphsignal/graphsignal

    Profile AI inference workloads (vLLM, SGLang, TensorRT-LLM, PyTorch, any GPU application) with the Graphsignal profiler and read the results from its local /signals JSON endpoint.

    257 GitHub stars~6.2k tokensUpdated 9 days ago
    AI & LLM EngineeringAuto-check passed

  • Quark Env Preflight

    amd/Quark

    Collect and normalize environment facts (OS, Python, GPU, CUDA/ROCm, container state) before Quark installation or PTQ planning.

    181 GitHub stars~1.4k tokensUpdated 9 days ago
    AI & LLM EngineeringAuto-check passed

  • Spark Environment Setup

    wshobson/agents

    Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13).

    40k GitHub stars~2k tokensUpdated 2 days ago
    AI & LLM EngineeringAuto-check passed

  • Ako4all

    TongmingLAIC/AKO4ALL

    Drive an agentic loop that iteratively optimizes a GPU kernel for maximum speedup.

    369 GitHub stars~4k tokensUpdated 22 days ago
    AI & LLM EngineeringAuto-check passed

  • Paddle Op Dev

    PaddlePaddle/Paddle

    PaddlePaddle (飞桨) C++ 算子开发指南。提供从 YAML 配置、InferMeta 函数、Kernel 实现、Python API 封装、单元测试到编译验证的完整算子开发流程指导。在以下场景使用此 skill：(1) 为 Paddle 框架新增 C++ 算子 (2) 修改或调试已有 Paddle 算子 (3) 编写算子的 YAML…

    24k GitHub stars~1.3k tokensUpdated 7 days ago
    AI & LLM EngineeringAuto-check passed

  • Embedded AI Deployment

    matlab/agent-skills-playground

    Deploy AI models to embedded hardware using MathWorks tools (MATLAB, Simulink, Embedded Coder).

    181 GitHub starsUsed in 1 repo~3.4k tokens
    AI & LLM EngineeringAuto-check passed

More from VectorSpaceLab/AREX-Skill

All 159 skills in this repo

  • Agent Lightning

    VectorSpaceLab/AREX-Skill

    Use this repo skill for Agent Lightning package tasks: authoring trainable agents, tracing rewards and spans, running LightningStore/Trainer loops, using agl CLI services, choosing examples, and…

    328 GitHub stars~1.4k tokensUpdated 1 mo ago
    Auto-check passed

  • Agent Tools

    VectorSpaceLab/AREX-Skill

    A skill your agent uses when configuring LiteLLM for MCP tools, A2A agents, Claude Code/Cursor agent gateway traffic, MCP auth/OAuth, tool permissions, semantic filtering, or agent-specific proxy…

    328 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed

  • Agents And Awel

    VectorSpaceLab/AREX-Skill

    Build and debug DB-GPT agents, tools, skills, teams, and AWEL workflows, including deterministic local DAG runs and HTTP-trigger topology without assuming an LLM, credential, or external service.

    328 GitHub stars~2.3k tokensUpdated 1 mo ago
    Auto-check passed

  • Agents And Middleware

    VectorSpaceLab/AREX-Skill

    Work on the actively maintained LangChain v1 agent package: initchatmodel, createagent, structured output, tools, middleware, embeddings initialization, provider routing, and agent runtime…

    328 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed

  • Agents Workflows

    VectorSpaceLab/AREX-Skill

    A skill your agent uses for giskard.agents async chat workflows, tools, prompt templates, structured outputs, retries, rate limiting, embeddings, and optional LiteLLM backend.

    328 GitHub stars~500 tokensUpdated 1 mo ago
    Auto-check passed

  • Alphafold3

    VectorSpaceLab/AREX-Skill

    A skill your agent uses for AlphaFold 3 input preparation, prediction command planning, output interpretation, and Python API inspection.

    328 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Categories

Questions about Torch Tensorrt

What does Torch Tensorrt do?

A skill your agent uses for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging…. Torch Tensorrt is an agent skill from VectorSpaceLab/AREX-Skill. Use this skill for Torch-TensorRT tasks: compiling PyTorch models with TensorRT, dynamic-shape/export workflows, runtime optimization, Triton/C++/distributed deployment, debugging unsupported ops, and maintaining source builds.

When should I use Torch Tensorrt?

Torch Tensorrt fits situations like: torch-TensorRT tasks: compiling PyTorch models with TensorRT; dynamic-shape/export workflows; runtime optimization; triton/C++/distributed deployment.

How do I install Torch Tensorrt in Claude Code?

Run `npx skills add VectorSpaceLab/AREX-Skill --skill torch-tensorrt -a claude-code`. Or copy the skill folder (skills/repositories/repo-skills/torch-tensorrt in VectorSpaceLab/AREX-Skill) into .claude/skills/torch-tensorrt in your project. Claude Code loads it when a task matches its description.

How do I install Torch Tensorrt in Codex?

Run `npx skills add VectorSpaceLab/AREX-Skill --skill torch-tensorrt -a codex`. Or copy the skill folder (skills/repositories/repo-skills/torch-tensorrt in VectorSpaceLab/AREX-Skill) into .agents/skills/torch-tensorrt in your project. Codex loads it when a task matches its description.

Can I use Torch Tensorrt in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add VectorSpaceLab/AREX-Skill --skill torch-tensorrt -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/torch-tensorrt, .gemini/skills/torch-tensorrt, .github/skills/torch-tensorrt and .opencode/skills/torch-tensorrt in your project.

What does Torch Tensorrt need to run?

Going by SKILL.md and its folder, Torch Tensorrt needs Python for the scripts in its folder and the command-line tools its instructions call (python). Our summary lists: Python 3.

Does Torch Tensorrt access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Torch Tensorrt safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Torch Tensorrt use?

Torch Tensorrt is published under the BSD-3-Clause licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Torch Tensorrt use?

About 1.5k tokens (SKILL.md is roughly 6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 5.4k tokens, read only when the agent opens those files.

What are the alternatives to Torch Tensorrt?

Skills that share tags, products or a category with Torch Tensorrt: Graphsignal (graphsignal/graphsignal, 257 stars), Quark Env Preflight (amd/Quark, 181 stars), Spark Environment Setup (wshobson/agents, 40k stars) and Ako4all (TongmingLAIC/AKO4ALL, 369 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Torch Tensorrt?

VectorSpaceLab (a GitHub organization) maintains it in VectorSpaceLab/AREX-Skill, which has 328 GitHub stars. The repository holds 159 skills in this directory. The repository was last updated on September 3, 2026.

Source: VectorSpaceLab/AREX-Skill on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.