← All skills
tao-run-deft-aoi
72 / 100 · Review before installing
Check before installing. Some of what it can do is expected: a coding skill usually runs build and test commands.
By NVIDIA/skills (GitHub) · coding · 3,933 stars · scanned at commit 2630133
Run the full DEFT AOI improvement loop for NVIDIA TAO VisualChangeNet / ChangeNet PCB inspection models: baseline evaluate, RCA, Cosmos AnomalyGen / AMP synthetic defects, k-NN mining, retraining, and deployment gating against a customer-defined primary metric and optional constraints. Use only when the request identifies an AOI / automated-optical-inspection, PCB-defect, VisualChangeNet, or ChangeNet workflow. Supports air-gapped/offline runs with pre-staged assets. Never infer AOI from generi…
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
This is disclosed by the author in the skill's SKILL.md or README. Expected for a coding skill.
Found in 22 places.
Show 17 more
- skills/tao-run-deft-aoi/cards/10-post-train.md, line 15
- `FileNotFoundError: /home/...model_epoch...` (host path in `pretrained_model_path`) → `bash -c 'sed -i -E "s|(pretrained_model_path:) *$WS|\1 /data/workspace| - skills/tao-run-deft-aoi/cards/10-post-train.md, line 21
bash -c 'set -e; T=$RD/$ITER/train; SEL=$($DPY -c " - skills/tao-run-deft-aoi/cards/10-post-train.md, line 41
bash -c 'set -e; SPEC=$RD/specs/$([ "$ITER" = baseline ] && echo baseline_spec.yaml || echo ${ITER}_spec.yaml); T=$RD/$ITER/train; CKB=$(basename $(ls $T/model_ - skills/tao-run-deft-aoi/cards/20-evaluate.md, line 9
bash -c 'ls $RD/$ITER/inference/best_val/inference/inference.csv $RD/$ITER/inference/latest/inference/inference.csv 2>/dev/null | wc -l' - skills/tao-run-deft-aoi/cards/20-evaluate.md, line 19
bash -c 'set -e; for N in best_val latest; do $DPY $SKILL_ROOT/scripts/analyze_kpi.py $RD/$ITER/inference/$N/inference/inference.csv --output-dir $RD/$ITER/infe - skills/tao-run-deft-aoi/cards/20-evaluate.md, line 24
bash -c 'set -e; W=$($DPY -c "import json;a=json.load(open(\"$RD/$ITER/inference/best_val_eval/metric_result.json\"));b=json.load(open(\"$RD/$ITER/inference/lat - skills/tao-run-deft-aoi/cards/30-post-evaluate.md, line 18
bash -c 'set -e; OUT=$RD/$ITER/rca_results/$(date +%s); mkdir -p $OUT; WIN=$($DPY -c "import json;print(json.load(open(\"$RD/deft_state.json\"))[\"iterations\"] - skills/tao-run-deft-aoi/cards/30-post-evaluate.md, line 24
bash -c 'set -e; OUT=$(ls -td $RD/$ITER/rca_results/*/ | head -1); TH=$($DPY -c "import json;print(json.load(open(\"$RD/deft_state.json\"))[\"iterations\"][\"$I - skills/tao-run-deft-aoi/cards/40-routing.md, line 16
bash -c 'set -e; OUT=$RD/$ITER/routing_results/$(date +%s); mkdir -p $OUT; $DPY -c " - skills/tao-run-deft-aoi/cards/40-routing.md, line 37
bash -c 'set -e; OUT=$(ls -td $RD/$ITER/routing_results/*/ | head -1); $DPY $SKILL_ROOT/scripts/commit_stage.py --duration-sec $(( $(date +%s) - STAGE_T0 + 1 )) - skills/tao-run-deft-aoi/cards/50-mining.md, line 11
bash -c '$DPY -c "import json,pandas as pd;st=json.load(open(\"$RD/deft_state.json\"));p=st[\"iterations\"][\"$ITER\"][\"routing_mining_parquet\"];print(\"ROWS= - skills/tao-run-deft-aoi/cards/50-mining.md, line 20
bash -c 'set -e; MF=$RD/$ITER/mining_filter; mkdir -p $MF/pkg_specs/{embedding,tmm}; TP=$($DPY -c "import json;print(json.load(open(\"$RD/deft_state.json\"))[\" - skills/tao-run-deft-aoi/cards/50-mining.md, line 45
bash -c 'set -e; MF=$RD/$ITER/mining_filter; K=$($DPY -c "import json;print(json.load(open(\"$RD/$ITER/mining_filter/mining_history_summary.json\"))[\"selected_ - skills/tao-run-deft-aoi/cards/60-merge-train.md, line 12
bash -c 'set -e; $DPY -c " - skills/tao-run-deft-aoi/cards/60-merge-train.md, line 50
bash -c 'set -e; SPEC=$RD/specs/${ITER}_spec.yaml; cp $RD/specs/baseline_spec.yaml $SPEC; PREV=$($DPY -c "import json;st=json.load(open(\"$RD/deft_state.json\") - skills/tao-run-deft-aoi/eval.slow-manual.config, line 16
"prompt": "Run the full DEFT Loop end-to-end as a smoke test on subset data with AnomalyGen + Data Mining.\n\n## Plugin Installation\nThis eval depends on the ` - skills/tao-run-deft-aoi/scripts/deft_exec.py, line 159
return subprocess.run(command, env=environment, check=False).returncode
and 2 more places
Changes files outside its own folder
The author does not mention this.
Found in 2 places.
Downloads and installs other software
The author does not mention this.
Found in 1 place.
Safer alternatives
Low-risk skills in the same category (coding) that look related. Check that one does what you need.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'tao-run-deft-aoi' ] || [ -L ~/.agents/skills/'tao-run-deft-aoi' ]; then echo 'A skill folder named tao-run-deft-aoi already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/NVIDIA/skills' "$d" && git -C "$d" checkout --quiet '2630133f22e357f2005886567a39eddcc1b520e2' -- 'skills/tao-run-deft-aoi' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/tao-run-deft-aoi' ~/.agents/skills/'tao-run-deft-aoi' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'tao-run-deft-aoi' ~/.claude/skills/'tao-run-deft-aoi' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for NVIDIA/skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/NVIDIA/skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-11 | 2630133 | 72 |
| 2026-11-08 | fce3cca | 72 |
| 2026-11-07 | fce3cca | 72 |