# Production Checklist

## Factory build

- [ ] Install Rust toolchain on factory PC
- [ ] Run `factory/powershell/Build-RustBinaries.ps1`
- [ ] Confirm `bin/JackAILocald.exe` exists
- [ ] Install or copy Ollama binary
- [ ] Confirm `backends/ollama/windows/ollama.exe` exists
- [ ] Install or copy llama.cpp binary
- [ ] Confirm `backends/llama.cpp/windows/llama-server.exe` exists
- [ ] Set `OLLAMA_MODELS` to USB `models/ollama`
- [ ] Pull required Ollama models into USB
- [ ] Add GGUF models if llama.cpp fallback is enabled
- [ ] Generate manifest
- [ ] Run Windows preflight
- [ ] Run first-launch from USB on clean Windows machine
- [ ] Run airplane-mode test
- [ ] Run benchmark demo
- [ ] Confirm no model download occurs during customer launch

## Security

- [ ] Remove placeholder files before release
- [ ] Replace update warning hook with signature verification
- [ ] Sign release artifacts
- [ ] Scan binaries with Microsoft Defender and VirusTotal-compatible workflow
- [ ] Document realistic host traces
- [ ] Avoid "leaves no trace" claim
- [ ] Bind all local services to 127.0.0.1 only
- [ ] Add process crash recovery
- [ ] Add request size limit
- [ ] Add timeout limit

## Linux boot edition

- [ ] Flash Linux image to SSD/USB
- [ ] Install Linux runtime
- [ ] Install Linux backends
- [ ] Configure kiosk auto-start
- [ ] Test boot on Dell/Lenovo/HP/ASUS at minimum
- [ ] Test Secure Boot behavior
- [ ] Test no-network mode
- [ ] Test GPU/CPU fallback