tao-run-deft-aoi-cosmos3 agent skill by NVIDIA
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.
Source and license
| Publisher | NVIDIA |
|---|---|
| Repository | NVIDIA/skills |
| License | Apache-2.0 AND CC-BY-4.0 |
| Snapshot | 2026-10-22 |
| Last repository commit | 2026-10-17 |
| Instruction size | 2,394 words; 335 lines |
| Bundled files | 61 |
Suede Skill Check: 92/100 · A
Report snapshot · · graded-v2
The rubric totals 100 points; published scores are capped at 99. Security scan: not scanned.
Source revision: 4fc9cf6cec78. Report version fc3b328ae445. Badge URLs follow the latest listing; save the report to keep this version.
Download report · Request a fresh review
What to review next
- Description length: 3 of 5 points. 5 for 150–600 characters; 3 for 100–900; 1 for 50–1,024.
Add declarations and bundled files when they describe the skill’s actual needs. Update the source, then request a fresh review with the changed commit or PR.
| Component | Points earned | Available |
|---|---|---|
| Structural checks | 56 | 56 |
| Description length | 3 | 5 |
| Boundary clause | 4 | 4 |
| Example request | 3 | 3 |
| Instruction depth | 7 | 7 |
| Reference files | 3 | 3 |
| Bundled scripts | 2 | 2 |
| Recent maintenance | 9 | 9 |
| Curated-list citations | 1 | 7 |
| Tool permissions declared | 2 | 2 |
| Compatibility declared | 2 | 2 |
The published score is capped at 99. It measures the published files, documentation and maintenance signals. Security scan: not scanned. This assessment does not include a runtime test or a full installation-dependency review.
Install tao-run-deft-aoi-cosmos3
The skills CLI works with Claude Code, Codex, Cursor and other agents; the other routes copy the folder by hand.
npx skills add https://github.com/NVIDIA/skills/tree/main/skills/tao-run-deft-aoi-cosmos3tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp/repo"
mkdir -p .claude/skills
cp -R "$tmp/repo/"skills/tao-run-deft-aoi-cosmos3 .claude/skills/
rm -rf "$tmp"tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp/repo"
mkdir -p .agents/skills
cp -R "$tmp/repo/"skills/tao-run-deft-aoi-cosmos3 .agents/skills/
rm -rf "$tmp"For a user-level install, copy the folder to ~/.claude/skills/ (Claude Code) or ~/.agents/skills/ (Codex). Inside Codex, $skill-installer can also fetch a skill from GitHub. New to skills? Read How to install agent skills in Claude Code, Codex and Cursor.
NVIDIA/skills is third-party code. Read SKILL.md and any bundled scripts before you install it.
Two badges for tao-run-deft-aoi-cosmos3
Choose the numeric badge, the letter badge, or both. Each links to this skill’s Suede AI listing; either valid linked badge can qualify for the feature offer.
[](https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/)
[](https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/)<a href="https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/"><img src="https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/score.svg" alt="Suede Skill Check: 92/100" width="160" height="175"></a>
<a href="https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/"><img src="https://suedeai.directory/nvidia/skills/tao-run-deft-aoi-cosmos3/badge.svg" alt="Suede Skill Check: Grade A" width="160" height="175"></a>Suede Certified on the shield refers to this published structural quality check. Security scan: not scanned. Read the methodology.
Limited-time feature offer: Display either quality badge in your repository or skill README, linked to your listing, then request verification. Suede AI checks the rendered README at a recorded commit. Eligible skills scoring 80–99 with a verified badge can receive a featured placement while the offer is open. Request badge verification.
Publishers: read How to get your agent skill found in search and cited by AI assistants.