Agent skill · nvidia
tao-run-deft-aoi-cosmos3
Run the disk-backed DEFT AOI improvement loop for NVIDIA Cosmos Reason 3 / Cosmos3 models, using Nano by default and Edge or Super when explicitly requested: evaluate the base model on Proxy and frozen Benchmark splits, mine real image pairs from Proxy gaps, generate AnomalyGen synthetic NG pairs, assemble a per-iteration Train JSON from both producers, train with Cosmos Framework LoRA SFT, and repeat through the selected platform's submit/status/logs/cancel contract. This migration supports bare labels only: the assistant response must be exactly OK or NG. Use for "run Cosmos3 DEFT AOI", "CR3 AOI loop", or "improve Cosmos3 PCB inspection with bare OK/NG"; do not use for rich/reasoning annotation, one-off Cosmos training, or generic anomaly generation.
What it needs
About 10k tokens when loaded.
What this skill does
tao-run-deft-aoi-cosmos3 Installation Install this application as part of the full TAO skill-bank root, not as only the companion skill folders: a usable install places skills/, scripts/, templates/, and versions.yaml under the same TAOSKILLBANKPATH. Any install that ships only the skill folders, for example an agent plugin or skills-only install, must also provide the bank-level scripts, templates, and versions.yaml and point TAOSKILLBANKPATH at their common root before running scripts/resolvetaomodel.py. Run bundled validation with the skill Python so dependencies match runtime: PYTHON=$(bash scripts/deftpython.sh); "$PYTHON" -m unittest tests.testcosmos3bare. Resolve network mode first. Missing air-gap imports are a hard stop; network-enabled setup lives only in references/network-bootstrap.md. Execution Contract Treat a run as a disk-backed state machine. 1. Preserve every explicit user value and show the source of each effective value (user, spec, or default) in the Pre-Flight Summary. 2. Ask which installed platform to use. Do not select Docker, SLURM, Kubernetes, Brev, virtualenv, or an external platform by default. 3. Resolve network mode, then read exactly one of references/air-gap.md or references/network-bootstrap.md. Run the selected platform skill's Preflight and stop on a missing system/native-CLI prerequisite. 4. Before any mutation or launch, invoke tao-launch-workflow and show its launch review plus this skill's Pre-Flight Summary. Wait for one explicit approval. 5. After approval, set PYTHON=$(bash scripts/deftpython.sh) and initialize ${RESULTSDIR}/deftstate.json once with "$PYTHON" scripts/initdeftstate.py. Require --base-model-path to name the prepared Qwen3-VL PTM, freeze the image with --framework-image-digest, and pass its container URI with --framework-container. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills nvidia/tao-run-deft-aoi-cosmos3