菜单
Use before running or interpreting NPA tests, lint checks, or validation reports.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when ingesting, validating, curating, or querying production sensor data as a versioned dataset-of-record, or wiring the dataset-ingest-curate workflow.
Use when generating adversarial scenarios via RL, ranking mined failures of a policy-under-test, or wiring the adversarial-scenario-hardening workflow.
Use when navigating the Cosmos3 integration in NPA or locating upstream Cosmos3 framework files, defaults, scripts, configs, recipes, and docs.
Use when Cosmos3 setup, fetch, inference, CUDA, uv, Docker, Hugging Face, GitHub, NGC, or checkpoint staging fails in NPA or in an upstream Cosmos framework checkout.
Use when setting up Cosmos3 access through NPA, checking source or Hugging Face reachability, staging the public Cosmos3 framework and checkpoint cache, or deciding which NPA workflow to use before inference.
Use for Claude Code reviews that need robotics, simulation, GPU-routing, sim-to-real, or BDD100K pipeline domain context.
| name | testing-conventions |
| description | Use before running or interpreting NPA tests, lint checks, or validation reports. |
Use the repository virtualenv. Never use bare python; use npa/.venv/bin/python.
Correct command from repo root:
npa/.venv/bin/python -m pytest npa/tests/ --ignore=npa/tests/e2e --timeout=120 -q
Correct command from inside npa/:
cd npa
.venv/bin/python -m pytest tests/ --ignore=tests/e2e --timeout=120 -q
ruff is available in the venv:
npa/.venv/bin/python -m ruff check <files>
tests/smoke/test_cosmos_serverless_smoke.py has 5 tests gated by NPA_COSMOS_SERVERLESS_SMOKE=1; they fail with Unable to list Nebius VPC networks for project project-smoke: unsupported.NPA_INTEGRATION_E2E=1 and excluded from standard runs with --ignore=npa/tests/e2e.e2e_pipeline pytest marker.token_factory_e2e marker (in npa/tests/e2e/test_token_factory_e2e.py). They self-skip without a real NEBIUS_TOKEN_FACTORY_KEY; the marker is in conftest _LIVE_MARKERS so the key is not scrubbed. Run with NEBIUS_TOKEN_FACTORY_KEY=... pytest npa/tests/e2e/test_token_factory_e2e.py.Expected baseline: 1242+ passed, 21 skipped, 1 xpassed, 0 failures.
Use evidence-based convergence: report numeric pass counts and exact failure messages, not subjective assessment.
Tests must not hit real infrastructure. Mock SSH, S3, Nebius APIs, GPUs, and network calls at the call site. CLI tests use typer.testing.CliRunner against npa.cli.main:app.