NVIDIA-NeMo/Gym is a library and infrastructure for evaluating and improving models and agents inside environments, where each environment defines tasks, agent interaction, verification, and execution state. It is for teams running reproducible evaluations or training at scale across settings such as code execution, tool calling, and sandboxes, and the catalogue entries provide skills and instructions for working with it.
Borrowing it
Nothing to install: this file belongs to NVIDIA-NeMo/Gym. 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-NeMo/Gym/main/.agents/skills/run-and-babysit-eval/SKILL.mdgit clone --depth 1 https://github.com/NVIDIA-NeMo/GymWrote 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-nemo/gym/run-and-babysit-eval)<a href="https://agentmods.dev/skills/nvidia-nemo/gym/run-and-babysit-eval"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/gym/run-and-babysit-eval/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-nemo/gym/run-and-babysit-eval"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/gym/run-and-babysit-eval.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 | $0.00190 | $0.09597 |
| Opus 5.5 | $0.00076 | $0.03839 |
| Sonnet 5.5 | $0.00038 | $0.01919 |
| Haiku 4.5 | $0.00019 | $0.00960 |
Grade C, and why
run-and-babysit-eval scanned grade C with 2 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 today.
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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
Without `uv`, install it with `curl -LsSf https://astral.sh/uv/install.sh | sh` and Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Without `uv`, install it with `curl -LsSf https://astral.sh/uv/install.sh | sh` and How it starts
The opening of the file, as written. The whole thing — 430 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run a Gym Evaluation to Its Result
The Gym pages are the reference; when they and this skill disagree, the pages win:
Configuration,
Resumable Runs,
Job Records and
Resume. For servers you started yourself
(gym env start, then gym eval run --no-serve) and for reading health while you build a benchmark or
a harness, follow add-benchmark and add-harness.
The page links point into a full Gym checkout (fern/...); a copy of this skill installed on its own does not
contain those pages, so read them in the Gym repository.
How to act
- Submit pinned and resumable (section 1), a small smoke first.
- Inspect, never guess (section 3): read
decision.actionandgate_task.status. - Do the one action the table names (section 4): run the same command without
--inspect, or add the one option the row names, with its reason. A decision with no row, or a STOP you cannot explain: stop and report it. - Never use plain
gym eval resubmit, edit a run's files, or regenerate a saved answer. - Done only when the inspect says
SUCCESS(section 6). Report the health, the corrections of section 7 and every operator action with its reason.
1. Submit
The gym commands below need the gym CLI on the host you submit from. Install it at the commit in the
config's driver.gym_install.ref, in a directory that persists (not /tmp), with uv on PATH:
SHA=<the full 40-character commit>
git clone -q https://github.com/NVIDIA-NeMo/Gym.git ~/gym-${SHA:0:12} && cd ~/gym-${SHA:0:12} && git checkout -q $SHA && uv venv --seed .venv && . .venv/bin/activate && uv pip install -e ".[mlflow]" && gym --help
What ships with it
2 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.
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.
- today Changed · -30 lines · -12 tokens per session cc5d8d02590a
- yesterday First seen · 460 lines · 202 tokens per session scan C 0f6926c19d75
run-and-babysit-eval is a skill published in the GitHub repository NVIDIA-NeMo/Gym (1,320 stars, last pushed today), licensed Apache-2.0. It adds 190 tokens to every session and 9,597 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-05.
run-and-babysit-eval compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| run-and-babysit-evalNVIDIA-NeMo/Gym | Skill | 1.3k | today | Apache-2.0 | C | 190 | Codex |
| model_finetuningvuralserhat86/antigravity-agentic-skills | Skill | 43 | 10mo ago | none | A | 67 | Claude Code, Codex |
| model_finetuningDonggangChen/antigravity-agentic-skills | Skill | 2 | 10mo ago | none | A | 67 | Claude Code, Codex |
| foundry-config-setupmicrosoft/agent-framework | Skill | 15k | yesterday | MIT | A | 65 | Claude Code, Codex |
| rowboat-spacesrowboatlabs/rowboat | Skill | 18k | yesterday | Apache-2.0 | A | 60 | Claude Code, Codex |
| create-tech-designhumanlayer/humanlayer | Skill | 13k | yesterday | Apache-2.0 | A | 16 | Claude Code, Codex |
| deploy-docker-composeomnigent-ai/omnigent | Skill | 11k | today | Apache-2.0 | A | 84 | Claude Code, Codex |