Vanilla WAA bootstrap automation

README.md

@@ -827,6 +827,45 @@ uv run python -m openadapt_ml.benchmarks.cli vm monitor --auto-shutdown-hours 2
 
 For complete VM management commands and Azure setup instructions, see [`CLAUDE.md`](CLAUDE.md) and [`docs/azure_waa_setup.md`](docs/azure_waa_setup.md).
 
+### 13.5 Screenshot Capture Tool
+
+Capture screenshots of dashboards and VMs for documentation and PR purposes:
+
+```bash
+# Capture all available targets
+uv run python -m openadapt_ml.benchmarks.cli screenshot
+
+# List available targets
+uv run python -m openadapt_ml.benchmarks.cli screenshot --list
+
+# Capture specific targets
+uv run python -m openadapt_ml.benchmarks.cli screenshot --target terminal
+uv run python -m openadapt_ml.benchmarks.cli screenshot --target azure-ops --target vnc
+
+# Custom output directory
+uv run python -m openadapt_ml.benchmarks.cli screenshot --output /path/to/screenshots
+
+# Without timestamp in filename
+uv run python -m openadapt_ml.benchmarks.cli screenshot --target terminal --no-timestamp
+```
+
+**Available targets:**
+
+| Target | Description |
+|--------|-------------|
+| `azure-ops` | Azure ops dashboard (localhost:8765) |
+| `vnc` | VNC viewer (localhost:8006) - Windows VM |
+| `terminal` | VM monitor terminal output (mock mode) |
+| `terminal-live` | VM monitor terminal output (live, requires running VM) |
+| `training` | Training dashboard (localhost:8080) |
+| `vm-screen` | Windows VM screen capture via QEMU |
+
+**Notes:**
+- Terminal screenshots use PIL to render terminal output as PNG images
+- Web page screenshots work best with playwright installed (`uv add playwright && playwright install chromium`)
+- On macOS, interactive capture using `screencapture` is available as a fallback
+- Screenshots are saved to `docs/screenshots/` by default with timestamps
+
 ---
 
 ## 14. Limitations & Notes

deprecated/Dockerfile.simple

@@ -0,0 +1,26 @@
+FROM dockurr/windows:latest
+
+RUN apt-get update && apt-get install -y fuse dos2unix wget curl python3 python3-pip git && rm -rf /var/lib/apt/lists/*
+
+ENV YRES="900"
+ENV XRES="1440"
+ENV RAM_SIZE="8G"
+ENV CPU_CORES="8"
+ENV VERSION="11e"
+ENV DISK_SIZE="30G"
+ENV ARGUMENTS="-qmp tcp:0.0.0.0:7200,server,nowait"
+
+COPY src/win-arena-container/client /client
+COPY src/win-arena-container/vm/setup /oem
+COPY src/win-arena-container/entry.sh /entry.sh
+COPY src/win-arena-container/entry_setup.sh /entry_setup.sh
+COPY src/win-arena-container/start_client.sh /start_client.sh
+COPY src/win-arena-container/start_vm.sh /start_vm.sh
+
+RUN pip3 install --no-cache-dir -r /client/requirements.txt 2>/dev/null || true
+
+RUN find / -maxdepth 3 -type f -name "*.sh" -exec dos2unix {} \; 2>/dev/null; chmod +x /*.sh 2>/dev/null || true
+
+RUN sed -i "s|20\.20\.20\.21|172.30.0.2|g" /entry_setup.sh /entry.sh /start_client.sh 2>/dev/null || true
+
+ENTRYPOINT ["/bin/bash", "-c"]

deprecated/README.md

@@ -0,0 +1,7 @@
+# Deprecated WAA Legacy Materials
+
+This folder contains legacy WAA automation files and documents that are no longer
+part of the vanilla WAA workflow. They are retained for review only.
+
+Use `docs/waa_vanilla_automation.md` and the scripts in `scripts/` for the
+current vanilla automation flow.

deprecated/docs/WAA_ACR_DESIGN.md

@@ -0,0 +1,87 @@
+# WAA ACR Design (Unattended + Vanilla)
+
+## Goals
+- Make WAA image pulls reliable (no Docker Hub throttling/timeouts).
+- Preserve unattended Windows install (no license prompts).
+- Use existing CLI/scripts wherever possible.
+
+## Constraints
+- Windows install must be fully unattended (VERSION=11e + OEM Azure mode).
+- Prefer vanilla WAA components; no dev-mode UNC paths.
+- Avoid custom tooling if existing commands cover the flow.
+
+## Proposed ACR Naming
+ACR names must be globally unique and <= 50 chars. Use a deterministic pattern tied to the subscription and region:
+
+```
+openadapt-evals-<region>-<suffix>
+```
+
+Suggested suffix: last 6 chars of the Azure subscription ID.
+
+Example (eastus + sub id ...1234ab):
+
+```
+openadapt-evals-eastus-1234ab
+```
+
+If name is taken, append `-01`, `-02`, etc.
+
+## Implementation Plan
+
+### 1) Create ACR + import WinArena image
+Use the existing helper script in `openadapt-evals`:
+
+```bash
+cd /Users/abrichr/oa/src/openadapt-evals
+./scripts/setup_acr.sh \
+  --acr-name openadapt-evals-eastus-1234ab \
+  --resource-group openadapt-agents \
+  --workspace openadapt-ml \
+  --location eastus
+```
+
+This script:
+- Creates the registry (Basic tier).
+- Imports `docker.io/windowsarena/winarena:latest`.
+- Grants `AcrPull` to the Azure ML workspace identity.
+
+### 2) Use ACR image for Azure ML runs
+No new code needed; use the existing config/env support:
+
+```bash
+export AZURE_DOCKER_IMAGE="openadapt-evals-eastus-1234ab.azurecr.io/winarena:latest"
+```
+
+Then run Azure evals as usual:
+
+```bash
+uv run python -m openadapt_evals.benchmarks.cli azure --workers 1 --task-ids notepad_1 --waa-path /path/to/WAA
+```
+
+### 3) Use ACR image for dedicated VM builds
+The VM flow already supports ACR via existing CLI commands:
+
+```bash
+uv run python -m openadapt_ml.benchmarks.cli vm pull-image --acr openadapt-evals-eastus-1234ab
+```
+
+When building the custom `waa-auto` image on the VM, set:
+
+```bash
+export WAA_SOURCE_IMAGE="openadapt-evals-eastus-1234ab.azurecr.io/winarena:latest"
+uv run python -m openadapt_ml.benchmarks.cli vm prepare-windows
+```
+
+This uses the simplified Dockerfile (OEM Azure mode) and keeps installs unattended.
+
+## Verification Checklist
+- ACR import succeeded (`az acr repository show --name <acr> --repository winarena`).
+- Azure ML run logs show pulls from the ACR login server.
+- VM `prepare-windows` completes without product key prompts.
+- WAA `/probe` endpoint responds on port 5000 after boot.
+
+## Notes
+- The simplified Dockerfile copies OEM assets from the source image and uses `VERSION=11e` for unattended installs.
+- If Windows prompts for a product key, treat it as a regression and follow `docs/RECURRING_ISSUES.md`.
+- Keep Azure ML and ACR in the same region to avoid throttling and reduce pull time.