on-device-verification
100 / 100 · Low risk
Fine to try.
Prove a converted or quantized LiteRT model on the actual device via the CompiledModel API - confirm GPU residency, compare device output against the source model, and diagnose device-only failures such as silent CPU fallback, whole-graph compile ceilings, and fp16 range breaks. Use after conversion or quantization, when device output is wrong or NaN, when a clean graph fails to compile only on device, or when GPU and CPU outputs are suspiciously identical.
What this skill can do
We found nothing in the files we read that matches our rules. That does not prove the skill is safe: see the limits of static scanning.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'on-device-verification' ] || [ -L ~/.agents/skills/'on-device-verification' ]; then echo 'A skill folder named on-device-verification already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/google-ai-edge/litert-samples' "$d" && git -C "$d" checkout --quiet '163222b0a76d2e80f959c1fc1f1918420e3b1e4d' -- 'skills/on-device-verification' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/on-device-verification' ~/.agents/skills/'on-device-verification' && 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/'on-device-verification' ~/.claude/skills/'on-device-verification' (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
[](https://zalavijaysinh.github.io/skilltrust/repos/google-ai-edge/litert-samples/)
Score history
| Date | Version | Score |
|---|---|---|
| 2026-11-09 | 163222b | 100 |