Tao Run On Brev

NVIDIAOfficialCreated by NVIDIA Corporation

Skill
Tool integration
Cloud and DevOps
Apache-2.0
Medium risk

From the author

Run a TAO training/evaluation/inference container on an NVIDIA Brev GPU instance. Instance provisioning (create/search/stop/delete/login) is delegated to the official brev-cli agent skill or the Brev MCP server; this skill covers only the TAO-specific part — running the container over `brev exec` via the four-verb docker contract. Trigger phrases include "run on Brev", "Brev GPU instance", "TAO on Brev", "submit job to Brev".

— the description by NVIDIA Corporation

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill provides the bridge between NVIDIA TAO Toolkit workflows and NVIDIA Brev GPU instances. It acts as a specialized execution layer that uses the brev CLI to run TAO container operations. By leveraging the four-verb docker contract, it allows users to submit, monitor, and manage TAO jobs on remote GPU hardware. It is designed to work in tandem with official Brev provisioning skills, focusing strictly on the container execution lifecycle once an instance is ready.

What it does

Executes TAO Toolkit containers on remote NVIDIA Brev GPU instances via brev exec.
Manages the lifecycle of TAO jobs including submission, status checking, and cancellation.
Handles secure NGC authentication on remote instances using stdin piping to avoid credential exposure.
Supports pre-pulling of container images to optimize GPU billing time.
Integrates with TAO job recording systems to track job state and backend references.
Supports single-node multi-GPU configurations for training and inference.

When to use it

When you need to run TAO Toolkit training or inference tasks on on-demand GPU hardware.
When you have already provisioned a Brev instance and need to deploy a TAO container to it.
When you want to automate the execution of TAO jobs across ephemeral cloud GPU resources.
When you need to manage containerized TAO workloads while maintaining strict security for NGC credentials.

Advantages

Separates instance management from container execution for modularity.
Prevents credential leakage by avoiding command-line arguments for sensitive keys.
Optimizes cloud costs by separating image pulling from GPU-intensive job execution.
Provides a consistent interface for TAO jobs across different infrastructure backends.

Usage examples

“Run my TAO training job on the existing Brev instance named gpu-node-01.”: The agent will verify the instance readiness, authenticate with NGC, pull the required TAO image, and initiate the training container using the specified job parameters.
“Check the status of my TAO job on Brev.”: The agent will query the remote instance using docker inspect to report the current state of the containerized job.

Install

Download the skill as a ZIP:
https://mp-staging.autogineer.com/agent-skills/nvidia/tao-run-on-brev/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip tao-run-on-brev.zip -d .agents/skills

Documentation by NVIDIA Corporation

Brev — TAO execution glue

Standalone install? If this session was not initialized by the TAO skill bank plugin, run the tao-setup skill first (host preflight, credentials, cross-skill discovery).

NVIDIA Brev provides on-demand GPU instances (pre-loaded with NVIDIA drivers, CUDA, Docker, and the NVIDIA Container Toolkit). Brev is instance-based: you provision an instance, run commands on it over brev exec, and delete it when done.

This skill is deliberately thin. Provisioning and managing instances — create, search by GPU/price, start/stop, delete, login — is owned by NVIDIA Brev's own agent skill, not duplicated here. This skill covers only the TAO-specific part: running a TAO container on a reached instance through the four-verb docker contract, deferring the container-how to tao-run-on-docker over brev exec.

Provisioning: use the official Brev skill or MCP

NVIDIA Brev publishes an agent skill that manages instances in natural language ("create an A100 instance", "search for GPUs under $3/hr", "stop all my instances"). Install it once — it self-registers into your agent's skills dir and is discovered at runtime:

curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/scripts/install-agent-skill.sh | bash
# installs to ~/.claude/skills/brev-cli/ , ~/.codex/skills/brev-cli/ , ~/.agents/skills/brev-cli/

Or connect the Brev MCP server (https://docs.nvidia.com/brev/_mcp/server). Either one owns login/auth quirks, placement IDs, GPU search, and teardown flags. It does not cover container execution on the instance — that is this skill.

Preflight for this skill: the brev CLI is on PATH and logged in (headless: brev login --token "$BREV_API_TOKEN" before any other call), and you can reach a target instance — poll with a two-word command until it succeeds before issuing real work (a fresh instance reports RUNNING before sshd is up):

for i in $(seq 1 60); do brev exec <instance> "echo ok" 2>/dev/null | grep -qx ok && break; sleep 5; done
brev exec <instance> "echo ok" 2>/dev/null | grep -qx ok || { echo "instance not exec-ready"; exit 1; }

The probe must be two words, quoted as one argument. A single-token probe (brev exec <instance> -- true) passes even when every real command is broken, because brev exec [instance...] <command> treats only the LAST positional as the command — so a lone true lands in the right slot by accident while docker run ... does not. See brev exec argument form below.

Allow ≥ 600 s for the first brev exec on a new instance (SSH bring-up + first container pull); a 60–120 s wrapper timeout truncates startup and looks like a spurious exec failed.

Storage

No shared NFS/Lustre — storage tier B/C via tao-data-io: stage inputs from S3 to the instance's local disk (or fetch in-container) and upload results to S3 before deleting the instance. Instance-local ~/ persists across stop/start but not across delete/create, so the results upload must precede teardown.

Execution — the four verbs (a compound over Docker)

Brev is a compound consumer: submit reaches an instance, then defers the container-how to the four docker verbs (tao-run-on-docker) run over brev exec. It is not a symmetric peer — teardown must additionally delete the instance to stop billing. $BANK = ${TAO_SKILL_BANK_PATH}.

submit — reach an instance (provision/reuse via the official Brev skill or MCP; reuse an existing instance by its instance_id; wait for readiness, above). Lint the assembled command, open the record to mint $JOB_ID before launch, then run the docker submit verb inside the instance and mark RUNNING:

redact_secrets.py lint <<<"$REMOTE_CMD"     # no inline secrets; creds as -e VAR
JOB_ID=$("$BANK/scripts/tao_job_record.py" open \
  --platform brev --image "$IMG" \
  --network-arch "$ARCH" --action "$ACTION" \
  --storage-tier "$TIER" --results-root "$RESULTS_ROOT")
brev exec <instance> "docker inspect '$JOB_ID' >/dev/null 2>&1 && { echo '$JOB_ID already submitted'; exit 0; }; docker run -d --name '$JOB_ID' --label 'tao-job=$JOB_ID' ..."
"$BANK/scripts/tao_job_record.py" mark "$JOB_ID" --state RUNNING \
  --backend-ref "<instance>/$JOB_ID"       # instance is part of the ref: the
                                           # container is unreachable without it

status / logs — brev exec <instance> "docker inspect $JOB_ID" / brev exec <instance> "docker logs $JOB_ID", mapped to the vocab exactly as the docker verbs do. Recover <instance> from the record's backend-ref.

cancel / teardown — remove the container, then for an ephemeral instance delete it (stops billing), then mark the record. Never leave an ephemeral instance running:

brev exec <instance> "docker rm -f $JOB_ID"
brev delete <instance>                      # ephemeral instances only
"$BANK/scripts/tao_job_record.py" mark "$JOB_ID" --state CANCELED --source agent

`brev exec` argument form

The remote command is one argument. The CLI signature is brev exec [instance...] <command>: every positional except the last is an instance name, and -- only ends flag parsing — it does not group the words after it. So brev exec <inst> -- docker inspect "$JOB_ID" is read as instances <inst> docker inspect plus command "$JOB_ID", and fails with could not look up instance "docker" / ssh: illegal option -- - (exit 255) — an error that reads like an instance or SSH fault but is a syntax fault. Quote every remote command as a single string, exactly as brev exec --help shows.

NGC auth once per instance — never put NGC_KEY on argv (it lands in the remote process table); pipe it to --password-stdin:

IMG=nvcr.io/nvidia/tao/tao-toolkit:7.3.0-pyt  # versions-key: images.tao_toolkit.pyt

# NGC auth (one-time per instance) — value never on argv.
# Single-quoted locally so $NGC_KEY expands in the instance's shell; export it
# there first (or pipe it in from the local shell, if the instance has no copy).
brev exec <instance> 'printf %s "$NGC_KEY" | docker login nvcr.io -u "$oauthtoken" --password-stdin'

# Verify auth without reading ~/.docker/config.json. Failure before a successful
# login = not authenticated; failure after = the key's org lacks entitlement.
brev exec <instance> "docker manifest inspect $IMG >/dev/null && echo AUTH_OK || echo AUTH_FAIL"

# Pull BEFORE the GPU run. `docker run` would pull implicitly, but the instance
# bills from boot, so a multi-GB first-time TAO pull is billed GPU-idle time.
# Pulling as its own step also separates a pull failure (auth/entitlement) from
# a training failure in the logs.
brev exec <instance> "docker image inspect $IMG >/dev/null 2>&1 || docker pull $IMG"

# Run a TAO job (the docker `submit` verb, over brev exec)
brev exec <instance> "docker inspect '$JOB_ID' >/dev/null 2>&1 && { echo '$JOB_ID already submitted'; exit 0; }; docker run -d --name '$JOB_ID' --label 'tao-job=$JOB_ID' --gpus all -v ~/data:/data -e NGC_KEY '$IMG' visual_changenet train -e /data/spec.yaml"

Multi-GPU and multi-node

Multi-node is not supported on Brev — instance-based, no cross-instance coordination. Multi-GPU on a single instance is supported (up to 8× H100 / A100 / L40S); torchrun --nproc-per-node=N or PyTorch DDP work within the instance.

Files

ATTRIBUTION.mdAdded by AutoGineer915 B
BENCHMARK.md6.9 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md8.1 KiB
config/skillspector-baseline.yaml1.2 KiB
evals/evals.json771 B
references/skill_info.yaml439 B
skill-card.md4.5 KiB
skill.oms.sig5.0 KiB

LICENSE

Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.


                                 Apache License
                           Version 2.0, January 2004
                        http://www.apache.org/licenses/

   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

   1. Definitions.

      "License" shall mean the terms and conditions for use, reproduction,
      and distribution as defined by Sections 1 through 9 of this document.

      "Licensor" shall mean the copyright owner or entity authorized by
      the copyright owner that is granting the License.

      "Legal Entity" shall mean the union of the acting entity and all
      other entities that control, are controlled by, or are under common
      control with that entity. For the purposes of this definition,
      "control" means (i) the power, direct or indirect, to cause the
      direction or management of such entity, whether by contract or
      otherwise, or (ii) ownership of fifty percent (50%) or more of the
      outstanding shares, or (iii) beneficial ownership of such entity.

      "You" (or "Your") shall mean an individual or Legal Entity
      exercising permissions granted by this License.

      "Source" form shall mean the preferred form for making modifications,
      including but not limited to software source code, documentation
      source, and configuration files.

      "Object" form shall mean any form resulting from mechanical
      transformation or translation of a Source form, including but
      not limited to compiled object code, generated documentation,
      and conversions to other media types.

      "Work" shall mean the work of authorship, whether in Source or
      Object form, made available under the License, as indicated by a
      copyright notice that is included in or attached to the work
      (an example is provided in the Appendix below).

      "Derivative Works" shall mean any work, whether in Source or Object
      form, that is based on (or derived from) the Work and for which the
      editorial revisions, annotations, elaborations, or other modifications
      represent, as a whole, an original work of authorship. For the purposes
      of this License, Derivative Works shall not include works that remain
      separable from, or merely link (or bind by name) to the interfaces of,
      the Work and Derivative Works thereof.

      "Contribution" shall mean any work of authorship, including
      the original version of the Work and any modifications or additions
      to that Work or Derivative Works thereof, that is intentionally
      submitted to Licensor for inclusion in the Work by the copyright owner
      or by an individual or Legal Entity authorized to submit on behalf of
      the copyright owner. For the purposes of this definition, "submitted"
      means any form of electronic, verbal, or written communication sent
      to the Licensor or its representatives, including but not limited to
      communication on electronic mailing lists, source code control systems,
      and issue tracking systems that are managed by, or on behalf of, the
      Licensor for the purpose of discussing and improving the Work, but
      excluding communication that is conspicuously marked or otherwise
      designated in writing by the copyright owner as "Not a Contribution."

      "Contributor" shall mean Licensor and any individual or Legal Entity
      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,
      publicly display, publicly perform, sublicense, and distribute the
      Work and such Derivative Works in Source or Object form.

   3. Grant of Patent License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      (except as stated in this section) patent license to make, have made,
      use, offer to sell, sell, import, and otherwise transfer the Work,
      where such license applies only to those patent claims licensable
      by such Contributor that are necessarily infringed by their
      Contribution(s) alone or by combination of their Contribution(s)
      with the Work to which such Contribution(s) was submitted. If You
      institute patent litigation against any entity (including a
      cross-claim or counterclaim in a lawsuit) alleging that the Work
      or a Contribution incorporated within the Work constitutes direct
      or contributory patent infringement, then any patent licenses
      granted to You under this License for that Work shall terminate
      as of the date such litigation is filed.

   4. Redistribution. You may reproduce and distribute copies of the
      Work or Derivative Works thereof in any medium, with or without
      modifications, and in Source or Object form, provided that You
      meet the following conditions:

      (a) You must give any other recipients of the Work or
          Derivative Works a copy of this License; and

      (b) You must cause any modified files to carry prominent notices
          stating that You changed the files; and

      (c) You must retain, in the Source form of any Derivative Works
          that You distribute, all copyright, patent, trademark, and
          attribution notices from the Source form of the Work,
          excluding those notices that do not pertain to any part of
          the Derivative Works; and

      (d) If the Work includes a "NOTICE" text file as part of its
          distribution, then any Derivative Works that You distribute must
          include a readable copy of the attribution notices contained
          within such NOTICE file, excluding those notices that do not
          pertain to any part of the Derivative Works, in at least one
          of the following places: within a NOTICE text file distributed
          as part of the Derivative Works; within the Source form or
          documentation, if provided along with the Derivative Works; or,
          within a display generated by the Derivative Works, if and
          wherever such third-party notices normally appear. The contents
          of the NOTICE file are for informational purposes only and
          do not modify the License. You may add Your own attribution
          notices within Derivative Works that You distribute, alongside
          or as an addendum to the NOTICE text from the Work, provided
          that such additional attribution notices cannot be construed
          as modifying the License.

      You may add Your own copyright statement to Your modifications and
      may provide additional or different license terms and conditions
      for use, reproduction, or distribution of Your modifications, or
      for any such Derivative Works as a whole, provided Your use,
      reproduction, and distribution of the Work otherwise complies with
      the conditions stated in this License.

   5. Submission of Contributions. Unless You explicitly state otherwise,
      any Contribution intentionally submitted for inclusion in the Work
      by You to the Licensor shall be under the terms and conditions of
      this License, without any additional terms or conditions.
      Notwithstanding the above, nothing herein shall supersede or modify
      the terms of any separate license agreement you may have executed
      with Licensor regarding such Contributions.

   6. Trademarks. This License does not grant permission to use the trade
      names, trademarks, service marks, or product names of the Licensor,
      except as required for reasonable and customary use in describing the
      origin of the Work and reproducing the content of the NOTICE file.

   7. Disclaimer of Warranty. Unless required by applicable law or
      agreed to in writing, Licensor provides the Work (and each
      Contributor provides its Contributions) on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
      implied, including, without limitation, any warranties or conditions
      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
      PARTICULAR PURPOSE. You are solely responsible for determining the
      appropriateness of using or redistributing the Work and assume any
      risks associated with Your exercise of permissions under this License.

   8. Limitation of Liability. In no event and under no legal theory,
      whether in tort (including negligence), contract, or otherwise,
      unless required by applicable law (such as deliberate and grossly
      negligent acts) or agreed to in writing, shall any Contributor be
      liable to You for damages, including any direct, indirect, special,
      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

   9. Accepting Warranty or Additional Liability. While redistributing
      the Work or Derivative Works thereof, You may choose to offer,
      and charge a fee for, acceptance of support, warranty, indemnity,
      or other liability obligations and/or rights consistent with this
      License. However, in accepting such obligations, You may act only
      on Your own behalf and on Your sole responsibility, not on behalf
      of any other Contributor, and only if You agree to indemnify,
      defend, and hold each Contributor harmless for any liability
      incurred by, or claims asserted against, such Contributor by reason
      of your accepting any such warranty or additional liability.

   END OF TERMS AND CONDITIONS

   APPENDIX: How to apply the Apache License to your work.

      To apply the Apache License to your work, attach the following
      boilerplate notice, with the fields enclosed by brackets "[]"
      replaced with your own identifying information. (Don't include
      the brackets!)  The text should be enclosed in the appropriate
      comment syntax for the file format. We also recommend that a
      file or class name and description of purpose be included on the
      same "printed page" as the copyright notice for easier
      identification within third-party archives.

   Copyright [yyyy] [name of copyright owner]

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

       http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Declares allowed tools, Needs an MCP server, Needs network access, Needs Node.js

The author says: Requires the brev CLI (https://github.com/brevdev/brev-cli) and an active brev login. Instance provisioning is handled by the official brev-cli agent skill or the Brev MCP server.

Security

What our scanner noticed· Medium risk

Large base64-encoded blobs
References to credential files
Instructions that use the network
Downloads piped into a shell

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

NVIDIA Corporation

Publisher

NVIDIA (Official)

Source

NVIDIA/skills · skills/tao-run-on-brev · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE-APACHE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-08

AutoGineer is not affiliated with or endorsed by NVIDIA Corporation or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download tao-run-on-brev.zip