What is scenario-gen?

A tool for creating difficult test situations that try to make an AI-controlled policy fail. It can use reinforcement learning, where a system learns through repeated feedback, or a deterministic search that needs no GPU.

Generating adversarial scenarios, ranking previously mined failures, and connecting the workflow through its service, command-line interface, or software library.

It exposes failure cases that ordinary testing may miss, helping teams improve a policy and build regression tests for hard situations.

scenario-gen — Skill for Claude Code and Codex by nebius

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add nebius/nebius-physical-ai --skill scenario-gen

Clone the repo

git clone --depth 1 https://github.com/nebius/nebius-physical-ai

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/scenario-gen/github.svg)](https://agentmods.dev/skills/nebius/nebius-physical-ai/scenario-gen)

Your own site

<a href="https://agentmods.dev/skills/nebius/nebius-physical-ai/scenario-gen"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/scenario-gen/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/nebius/nebius-physical-ai/scenario-gen"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/scenario-gen.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 32 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,084 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

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	$0.00032	$0.01084
Opus 5	$0.00016	$0.00542
Sonnet 5	$0.00006	$0.00217
Haiku 4.5	$0.00003	$0.00108

Measured 4d ago against content hash bf6599bea42a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

scenario-gen 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 4d ago.

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.

skills/tools/scenario-gen/SKILL.md · 95 lines

How it starts

The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Scenario Gen (Adversarial Scenario Generation)

Adversarial scenario generation productizes the Isaac Lab RL capability as a first-class hard-case miner: an adversary perturbs the environment / other agents to maximize failures of a policy-under-test, surfacing hard scenarios for regression and hardening.

The adversary backend is pluggable. The intended production backend is an Isaac Lab RL adversary (reward = the policy-under-test's violation rate). The default backend is not RL — it is a deterministic, GPU-free heuristic search that acts as a functional scaffold/stand-in so the tool runs and is testable without a GPU. Plug in the real backend via adversary_backend.

Three-access pattern

Source of truth is the FastAPI service (npa/src/npa/workbench/scenario_gen/service.py). The CLI (npa/src/npa/cli/workbench/scenario_gen.py) and SDK (npa/src/npa/sdk/workbench/scenario_gen.py) are thin clients. Do not duplicate logic across layers.

Interfaces

CLI:

npa workbench scenario-gen generate --policy-uri <s3> --input-path <s3> --output-path <s3>
npa workbench scenario-gen rank --input-path <s3-manifest> --output-path <s3>
npa workbench scenario-gen status --run-id <id>
npa workbench scenario-gen system-info
npa workbench scenario-gen list

Endpoints: /health, /status, /system-info, /list, POST /generate, POST /rank.

API contract

POST /generate: given a policy-under-test checkpoint URI (--policy-uri) and a base task/scene config (--input-path), train an adversarial RL agent whose reward is the failure/violation of the policy-under-test, then emit a ranked adversarial set to --output-path. Output schema npa.scenario_gen.adversarial_set.v1 (S3 manifest + per-scenario configs and predicted failure metrics). Lineage (workflow run, input URIs, policy checkpoint, task) is threaded into every manifest.
POST /rank: score/rank a generated set by weighted failure severity + diversity; emits npa.scenario_gen.ranked_set.v1.

Read the full file on GitHub · 95 lines

Changes

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.

4d ago Changed bf6599bea42a
12d ago First seen · 95 lines · 32 tokens per session scan A 6475451a5ccc

Subscribe to this mod's changes

scenario-gen is a skill published in the GitHub repository nebius/nebius-physical-ai (29 stars, last pushed today), licensed Apache-2.0. It adds 32 tokens to every session and 1,084 once invoked, about $0.0002 per session on Opus 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.

Testing

research-engineer

An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.

davila7/claude-code-templates · 43 tokens

tika-eval-compare

Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".

apache/tika · 50 tokens

neuron-evaluation-engineer

Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…

neuron-core/neuron-ai · 77 tokens

scenario-gen