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.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/instructions/nvidia-nemo/gym/agents-md)<a href="https://agentmods.dev/instructions/nvidia-nemo/gym/agents-md"><img src="https://agentmods.dev/badge/instructions/nvidia-nemo/gym/agents-md/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/instructions/nvidia-nemo/gym/agents-md"><img src="https://agentmods.dev/badge/instructions/nvidia-nemo/gym/agents-md.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 | $0.05153 | $0.05153 |
| Opus 5.5 | $0.02061 | $0.02061 |
| Sonnet 5.5 | $0.01031 | $0.01031 |
| Haiku 4.5 | $0.00515 | $0.00515 |
Grade A, and why
Gym AGENTS.md 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 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.
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.
How it starts
The opening of the file, as written. The whole thing — 303 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Unified root instructions for AI coding assistants (Cursor, Claude, Codex, OpenCode, Copilot, and similar).
CLAUDE.md is a symlink to this file so Claude Code and other tools share one source of truth.
Humans: see Development Setup → Use of AI and LLM Tools (maintainer response policy) and Agent Skills.
Quality bar
- Prefer focused changes. Do not make unrelated "drive-by" edits. If a drive-by fix is worth keeping, open a separate issue or PR.
- Intentional synthetic scaling of environments is fine when scoped via an issue or focused PR; do not dump unreviewed bulk diffs.
- You (the human author) own every line submitted. Treat model output as untrusted until reviewed.
- For behavior-changing environment or agent work: run representative real smoke rollouts with a model and inspect agent and verifier behavior. Green unit tests alone are not enough. Metadata-only catalog or manifest changes and docs-only changes do not require model compute.
- Before opening a PR, run the local checks that mirror CI: tests (skip or N/A for docs-only),
pre-commit run --all-files, and DCO sign-off (git commit -s). Cryptographic-Ssigning is optional and not required. - AI-generated tests must assert real behavior; avoid vacuous pass-through tests.
- Prefer the vetted skills under
.agents/skills/(see Agent Skills). - For session lifecycle and routing terms (lock, fence, close retry window, saved reply, release, owner, episode session),
use the names in the Sessions section of
fern/versions/latest/pages/about/concepts/key-terminology.mdxand do not add synonyms. - Docs live under
fern/versions/latest/pages/. Bleeding-edge nav isfern/versions/main.yml. Seefern/README.mdand thenemo-gym-docsskill. - Do not introduce licenses incompatible with Apache-2.0. New source files need the standard NVIDIA SPDX header.
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 · +3 lines · +59 tokens per session 76cfed4742b9
- yesterday Changed · +26 lines · +559 tokens per session ffc73f0042b7
- 2d ago Changed · +13 lines · +324 tokens per session bf458832d7d0
- 3d ago Changed · +3 lines · +87 tokens per session 83e22971aca8
- 6d ago Changed · +13 lines · +266 tokens per session fb98e5eb0f54
- 10d ago Changed · +9 lines · +134 tokens per session e6e2805eaa85
- 16d ago Changed · +31 lines · +692 tokens per session 3acc3aea8dd7
- 29d ago Changed · +5 lines · +85 tokens per session c0d374d99036
- 1mo ago Changed · +35 lines · +658 tokens per session f9efdad44b78
- 2mo ago First seen · 165 lines · 2,289 tokens per session scan A 5556d6287ced
Gym AGENTS.md is an instructions file published in the GitHub repository NVIDIA-NeMo/Gym (1,320 stars, last pushed today), licensed Apache-2.0. It adds 5,153 tokens to every session, about $0.0206 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 2026-08-30.
Gym AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| Gym AGENTS.mdNVIDIA-NeMo/Gym | Instructions file | 1.3k | today | Apache-2.0 | A | 5,153 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |