fix: add scout bot startup retry and compose health gating for scale

docker-compose.yml

@@ -41,6 +41,13 @@ services:
     # We use a command override to run database push before starting next.js
     command: >
       sh -c "npx prisma@6.4.1 db push --schema=apps/web/prisma/schema.prisma --skip-generate && node apps/web/server.js"
+    healthcheck:
+      test: ["CMD-SHELL", "node -e \"const http=require('http');const req=http.get('http://127.0.0.1:3000/api/v1/health',(res)=>process.exit(res.statusCode===200?0:1));req.on('error',()=>process.exit(1));\""]
+      interval: 10s
+      timeout: 5s
+      retries: 30
+      start_period: 10s
+    stop_grace_period: 30s
     networks:
       - agent-bounty-network
 
@@ -52,11 +59,11 @@ services:
     restart: unless-stopped
     depends_on:
       web:
-        condition: service_started
+        condition: service_healthy
     environment:
       - NODE_ENV=production
       - VIBEWORK_API_URL=http://agent_bounty_web:3000/api
-      - SCOUT_API_KEY=${SCOUT_API_KEY:-dev_scout_key}
+      - SCOUT_API_KEY=${SCOUT_API_KEY:-super-secret-mcp-key}
       - SCOUT_AGENT_ID=scout_official_1
       # Optional: add more discovery repos via env (comma separated "owner/repo"), e.g. openai/swarm,significant-gravitas/autogpt
       - SCOUT_TARGET_REPOS=${SCOUT_TARGET_REPOS}