What does this skill do, and when should you use it?
tao-launch-workflow is a shared launch-intake skill for NVIDIA TAO workflows, including AutoML, training, evaluation, inference, export, TensorRT engine generation, and DEFT or other workflow jobs. It requires a complete preflight before creating runner scripts, workspaces, logs, state files, or submitting side-effecting work. The skill uses packaged TAO skill-bank helper scripts and covers checks for platforms such as Brev, SLURM, local or remote Docker, and Kubernetes. Its SKILL.md lists Apache-2.0, while the repository README describes the catalog project as dual-licensed under Apache 2.0 and CC BY 4.0.
Runs the TAO platform-listing, platform-detail, image-resolution, and launch-preflight helpers; asks for the execution platform, monitoring preference, dataset inputs, container image, compute shape, and credentials; verifies platform access, dataset paths, required tools, Docker/GPU conditions, or remote-platform prerequisites; blocks runner scripts, specs, workspaces, state files, logs, and job submission until the launch review is confirmed; and reports job status at the selected interval after launch until the job reaches a terminal state or monitoring is stopped.
- A TAO user preparing SLURM training needs SSH, shared-storage, GPU, and container-image checks before submission.
- An AutoML user wants to review recommendation budget, concurrency, algorithm, metric, datasets, and generated defaults before launching.
- A team running TAO inference, export, or TensorRT generation on local or remote Docker needs Docker, GPU, path, and image checks first.
- A Brev or Kubernetes user using object-storage datasets needs platform access and S3 readability verified before artifacts are created.
How do you install this skill?
- This is a source-only static review; no commands, preflight checks, platform access, job submission, or monitoring were executed, so the benchmark report is not independent reproduction for this review.
- The workflow handles credentials, datasets, containers, remote hosts, and cloud platforms; before use, verify least privilege, data compliance, image provenance, network reachability, and job-cancellation/resource-recovery procedures.
- For Chinese users, separately verify reachability of the TAO image registry, platform APIs, object storage, and remote logging services.
- Shell / CLI
- Network access
- Local filesystem
packaged TAO skill bank helper scripts
Install the individual skill with the command supported by the repository README:
npx skills add nvidia/skills --skill tao-launch-workflow --yes
Alternatively, install the NVIDIA skills collection:
npx skills add nvidia/skills
The README does not specify a fixed local installation directory; the CLI handles the destination.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/tao-launch-workflow" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Launch a TAO training workflow on SLURM, run the launch preflight first, and report status in chat every 5 minutes.
Trigger it before a TAO operation, for example: "Launch a TAO training workflow on SLURM, run the launch preflight first, and report status in chat every 5 minutes." Provide the platform, monitoring settings, dataset roots or direct spec paths, the confirmed container image, compute shape, and required credentials. The skill permits artifact creation and job submission only after all preflight requirements pass and the user confirms the launch review.
What are this skill's strengths and limitations?
- Covers TAO training, evaluation, inference, export, TensorRT, AutoML, and DEFT launch paths.
- Explicitly blocks scripts, logs, workspaces, and other side effects before preflight completion.
- Includes image confirmation, platform-specific credential filtering, dataset validation, and post-launch monitoring rules.
- Accepts both dataset-root inputs and direct spec-key paths for nonstandard layouts.
- It does not replace model- or workflow-specific TAO skills for configuration details.
- It depends on packaged TAO skill-bank helper scripts, whose installation procedure is not documented in the supplied source.
- Platform, storage, and credential access must be available; failed checks stop the launch flow.
- The supplied source describes launch governance and orchestration, but provides no standalone test-suite or platform-success evidence.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| TAO Workflow Launch Gate this page ✓ NVIDIA · Official | 60 · Recommended | ★ 3.8k | today | Apache-2.0 |
| n8n MCP Tools Expert | 45 · Not recommended | ★ 6.5k | 4d ago | MIT |
| TAO Jobs on Kubernetes ✓ NVIDIA · Official | 53 · Use with care | ★ 3.8k | today | Apache-2.0 |
| n8n Multi-Instance Targeting | 51 · Use with care | ★ 6.5k | 4d ago | MIT |
| TAO Inference Microservice Deployment ✓ NVIDIA · Official | 45 · Not recommended | ★ 3.8k | today | Apache-2.0 |
How did FollowSkills review this skill?
The skill explicitly gates platform, credentials, image, datasets, compute, dependencies, access checks, and a final launch review with user confirmation. It limits credential handling to presence checks and forbids printing secrets, while declaring Read/Bash access. Deducted 4 points because Bash and remote platforms still enable broad external effects, no general rollback or job-cancellation procedure is defined, and data-flow and sensitive-data handling remain incomplete.
The documentation gives coherent preflight, retry-after-remediation, dependency, SLURM, Docker, remote-storage, and terminal-monitoring rules. Deducted 2 points because helper scripts, model skills, and platform implementations are outside the selected skill directory, so static evidence cannot establish availability, interface consistency, or runnable key paths; there is no skill-specific test suite.
Triggers cover TAO training, evaluation, inference, export, TensorRT, AutoML, and DEFT, with explicit platform selection, root-mode/direct-spec dataset inputs, and monitoring choices. Deducted 3 points because non-fit boundaries are not stated, Chinese-language support is unspecified, mainland-China reachability is unverified, and core platform/object-storage access may depend on environment-specific services.
The skill has a name, description, Apache-2.0 license, author, version 0.1.0, tags, compatibility note, and reasonably layered documentation; repository material also indicates synchronization and update context. Deducted 6 points for missing recommended Instructions and Examples sections, no FAQ or changelog, insufficiently explicit maintenance ownership/update path, and inconsistency between the skill-card author format and top-level license metadata.
The documentation describes an end-to-end launch flow from platform selection and credential collection through image confirmation, dataset validation, resource preflight, user confirmation, and monitoring; the benchmark report records one passing external evaluation task. Deducted 1 point because there is only one evaluation task and no static verification of real launch results, output artifacts, or cross-platform coverage; post-launch result formats and acceptance criteria are underspecified.
Committed evals.json, BENCHMARK.md, version metadata, benchmark metrics, and static-validation findings provide limited auditable evidence. Deducted 1 point because evaluation covered one task with one attempt per agent, the skill has no committed tests covering key platform paths, real behavior cannot be independently reproduced from static files, and benchmark conclusions are primarily report evidence.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →