Integrate CarMax PR #57 with reviewed fixes and harder tasks

.assets-revision

@@ -26,3 +26,6 @@ site.recreation_gov: 00faedd20a96b4b10d5013a1c07ace9472d7ee80
 
 # BoardGameGeek HF PR #25: immutable reviewed bundle; preserve all other pins.
 site.boardgamegeek: 6737719a7eeb8d9b5fba77072d455e0d9afd2a93
+
+# CarMax original asset PR #15, immutable reviewed seed and stock imagery.
+site.carmax: 189b9cdcfd41a21b99ce4797f3bec152d8564a7f

.gitattributes

@@ -0,0 +1,5 @@
+# Force LF line endings for shell scripts so they execute in Linux containers
+# even when checked out on Windows.
+*.sh        text eol=lf
+*.bash      text eol=lf
+Dockerfile  text eol=lf

AGENTS.md

@@ -4,7 +4,7 @@ A coding agent (Claude Code, Cursor, Aider, Codex, ...) is reading this. Read on
 
 ## What it is
 
-37 Flask mirror websites (Amazon, GitHub, BBC News, ...) packaged into one Docker image, plus a control plane on `:8101` for resetting per-site state. Used as a deterministic offline environment for web-agent benchmarks. ~3 GB image.
+38 Flask mirror websites (Amazon, GitHub, BBC News, ...) packaged into one Docker image, plus a control plane on `:8101` for resetting per-site state. Used as a deterministic offline environment for web-agent benchmarks. ~3 GB image.
 
 Two repos:
 - **code** (this one) — Flask apps, control plane, scripts.
@@ -48,17 +48,17 @@ Inside the image, sites live at `/opt/WebSyn/<site>/`. The path predates the ren
 # fresh clone
 ./scripts/fetch_assets.sh                     # pulls assets from HF
 ./scripts/build.sh                            # docker build -t webharbor:dev .
-docker run -d -p 8101:8101 -p 40000-40036:40000-40036 webharbor:dev
+docker run -d -p 8101:8101 -p 40000-40037:40000-40037 webharbor:dev
 ```
 
 Or use the published image directly:
 
 ```bash
-docker run -d -p 8101:8101 -p 40000-40036:40000-40036 \
+docker run -d -p 8101:8101 -p 40000-40037:40000-40037 \
   battalion7244/webharbor:latest
 ```
 
-Sites are on `40000`-`40036` in the order declared by `SITES=( ... )` in `websyn_start.sh`. Control plane:
+Sites are on `40000`-`40037` in the order declared by `SITES=( ... )` in `websyn_start.sh`. Control plane:
 
 | Method | Path                | Purpose                                   |
 |--------|---------------------|-------------------------------------------|
@@ -136,13 +136,13 @@ python3 -m py_compile sites/<site>/app.py
 
 # 3. run on alt ports (don't collide with anything you already have running)
 docker run -d --rm --name wh-test \
-  -p 8201:8101 -p 41000-41036:40000-40036 webharbor:dev
+  -p 8201:8101 -p 41000-41037:40000-40037 webharbor:dev
 
 # 4. control plane healthy, all sites alive
 curl -s http://localhost:8201/health | python3 -m json.tool | head
 
 # 5. every site renders 200
-for p in $(seq 41000 41036); do
+for p in $(seq 41000 41037); do
   curl -so /dev/null -w "$p:%{http_code}\n" http://localhost:$p/
 done
 

Dockerfile

@@ -1,5 +1,5 @@
 # WebHarbor — slim, self-contained image.
-# 37 Flask mirror sites + control plane on :8101.
+# 38 Flask mirror sites + control plane on :8101.
 
 FROM python:3.12-slim-bookworm
 
@@ -125,6 +125,6 @@ print('AccuWeather seed DB generated at build time.')" && rm -rf /opt/WebSyn/acc
 # Upgrade the pinned Recreation.gov seed before it becomes the reset fixture.
 RUN cd /opt/WebSyn/recreation_gov && python3 migrate_seed.py
 
-EXPOSE 8101 40000-40036
+EXPOSE 8101 40000-40037
 
 CMD ["/opt/websyn_start.sh"]

README.md

@@ -36,18 +36,18 @@ WebHarbor takes a different approach. We leverage coding agent (e.g., Claude Cod
 - **Deep features unlocked** — carts, checkouts, accounts, all fully testable
 - **Evolving** — harder tasks drive richer mirrors; the environment grows with agents
 - **RL-ready** — sub-second database resets between rollouts
-- **Community-driven** — 37 sites today, scaling to 100+ together
+- **Community-driven** — 38 sites today, scaling to 100+ together
 
 ## 🚀 Quickstart
 
 Build this checkout to run its registered web environments (published image tags may have an older registry):
 
 ```bash
 ./scripts/build.sh webharbor:dev
-docker run -p 8101:8101 -p 40000-40036:40000-40036 webharbor:dev
+docker run -p 8101:8101 -p 40000-40037:40000-40037 webharbor:dev
 ```
 
-Then point your agent at `http://localhost:40000` through `http://localhost:40036` to explore 37 local mirrors of WebVoyager sites: `Allrecipes, Amazon, Apple, ArXiv, BBC News, Booking, GitHub, Google Flights, Google Maps, Google Search, Hugging Face, Wolfram Alpha, Cambridge Dictionary, Coursera, ESPN, Merriam-Webster, IKEA, Phys.org, Target, TED, Ohio State University, Rotten Tomatoes, Compass, Walmart Careers, FedEx, WebMD Doctor, Healthline, Kaggle, NVIDIA, UC Berkeley, B&H Photo, AccuWeather, GOV.UK, IMDb, NBA, Recreation.gov, and BoardGameGeek`.
+Then point your agent at `http://localhost:40000` through `http://localhost:40037` to explore 38 local mirrors of WebVoyager sites: `Allrecipes, Amazon, Apple, ArXiv, BBC News, Booking, GitHub, Google Flights, Google Maps, Google Search, Hugging Face, Wolfram Alpha, Cambridge Dictionary, Coursera, ESPN, Merriam-Webster, IKEA, Phys.org, Target, TED, Ohio State University, Rotten Tomatoes, Compass, Walmart Careers, FedEx, WebMD Doctor, Healthline, Kaggle, NVIDIA, UC Berkeley, B&H Photo, AccuWeather, GOV.UK, IMDb, NBA, Recreation.gov, BoardGameGeek, and CarMax`.
 
 For sub-second reset between rollouts, expose the control plane and call `/reset/<site>`:
 
@@ -66,10 +66,11 @@ git clone https://github.com/aiming-lab/WebHarbor && cd WebHarbor
 
 ### Site registry
 
-This checkout registers **37 sites**. NVIDIA remains at index 28, UC Berkeley
+This checkout registers **38 sites**. NVIDIA remains at index 28, UC Berkeley
 remains at index 29, B&H Photo remains at index 30, AccuWeather remains at index
 31, GOV.UK remains at index 32, IMDb remains at index 33 and NBA remains at
-index 34. Recreation.gov remains at index 35; BoardGameGeek is appended at index 36. Build the image from
+index 34. Recreation.gov remains at index 35; BoardGameGeek remains at index 36,
+and CarMax is appended at index 37. Build the image from
 this checkout to use this registry; publishing source does not update the
 published Docker image automatically.
 
@@ -84,17 +85,18 @@ published Docker image automatically.
 | NBA | 34 | 40034 | 48034 |
 | Recreation.gov | 35 | 40035 | 48035 |
 | BoardGameGeek | 36 | 40036 | 48036 |
+| CarMax | 37 | 40037 | 48037 |
 
 `websyn_start.sh`, `control_server.py`, the `Dockerfile` `EXPOSE` line and every
-site's `tasks.jsonl` `web` URL agree on 37 sites and `40000-40036`;
+site's `tasks.jsonl` `web` URL agree on 38 sites and `40000-40037`;
 `scripts/check_site_registry.py` (run by `scripts/check_assets.sh`) fails when they
 drift.
 
 After preparing the candidate assets and building `webharbor:dev`, the local
 review deployment uses:
 
 ```bash
-docker run -p 127.0.0.1:48080:8101 -p 127.0.0.1:48000-48036:40000-40036 webharbor:dev
+docker run -p 127.0.0.1:48080:8101 -p 127.0.0.1:48000-48037:40000-40037 webharbor:dev
 ```
 
 NVIDIA inherits the site contribution from @KaKituken

control_server.py

@@ -32,7 +32,7 @@
     'fedex', 'webmd_doctor', 'healthline', 'kaggle',
     'nvidia', 'berkeley', 'bh_photo', 'accuweather', 'gov_uk',
     'imdb',
-    'nba', 'recreation_gov', 'boardgamegeek',
+    'nba', 'recreation_gov', 'boardgamegeek', 'carmax',
 ]
 BASE_PORT = 40000
 WEBSYN_DIR = '/opt/WebSyn'

scripts/verify_carmax.sh

@@ -0,0 +1,43 @@
+#!/usr/bin/env bash
+# Build and verify all registered sites; reset CarMax byte-identically.
+# Uses an owned container only. Override ports/tag with CARMAX_VERIFY_*.
+set -euo pipefail
+cd "$(dirname "$0")/.."
+
+verify_container="wh-carmax-verify-$$"
+verify_image="${CARMAX_VERIFY_IMAGE:-webharbor:carmax-check}"
+verify_control="${CARMAX_VERIFY_CONTROL_PORT:-8201}"
+verify_low="${CARMAX_VERIFY_BASE_PORT:-41000}"
+verify_count=$(python3 -c 'import ast; m=ast.parse(open("control_server.py").read()); print(len(next(ast.literal_eval(n.value) for n in m.body if isinstance(n,ast.Assign) and any(isinstance(t,ast.Name) and t.id=="SITES" for t in n.targets))))')
+verify_high=$((verify_low + verify_count - 1))
+verify_internal_high=$((40000 + verify_count - 1))
+verify_started=0
+cleanup() {
+    if (( verify_started )); then docker stop "$verify_container" >/dev/null; fi
+}
+trap cleanup EXIT
+
+./scripts/build.sh "$verify_image"
+docker run -d --rm --name "$verify_container" \
+    -p "127.0.0.1:$verify_control:8101" \
+    -p "127.0.0.1:$verify_low-$verify_high:40000-$verify_internal_high" \
+    "$verify_image"
+verify_started=1
+verify_ready=0
+for ((attempt=0; attempt<120; attempt++)); do
+    if curl -fsS --max-time 5 "http://localhost:$verify_control/health" >/dev/null 2>&1; then
+        verify_ready=1
+        break
+    fi
+    sleep 1
+done
+if (( ! verify_ready )); then docker logs "$verify_container"; exit 1; fi
+for ((port=verify_low; port<=verify_high; port++)); do
+    code=$(curl -sS --max-time 15 -o /dev/null -w '%{http_code}' "http://localhost:$port/")
+    [[ "$code" == 200 ]] || { echo "Homepage $port returned $code"; exit 1; }
+done
+curl -fsS --max-time 30 -X POST "http://localhost:$verify_control/reset/carmax"
+verify_hashes=$(docker exec "$verify_container" md5sum \
+    /opt/WebSyn/carmax/instance/carmax.db /opt/WebSyn/carmax/instance_seed/carmax.db)
+[[ $(awk '{print $1}' <<< "$verify_hashes" | sort -u | wc -l) -eq 1 ]]
+echo "PASS: $verify_count homepages, healthy control plane, byte-identical CarMax reset."