Memory update: Feed Hunter live, email setup, control panel building

MEMORY.md

@@ -66,11 +66,19 @@ This is about having an inner life, not just responding.
   - Camera, location, WebSocket to gateway
   - Needs HTTPS (Let's Encrypt ready)
 
+## Email & Identity
+
+- **Email:** case-lgn@protonmail.com (credentials in .credentials/email.env)
+- D J set this up 2026-02-08 — big trust milestone
+- Used for API registrations, service signups
+
 ## Active Threads
 
+- **Feed Hunter:** ✅ Pipeline working, first sim running (Super Bowl 2026-02-08)
+- **Control Panel:** Building at localhost:8000 (accounts/API keys/services/budget)
 - **Sandbox buildout:** ✅ Complete (74 files, 37 tools)
 - **Inner life system:** ✅ Complete (7 tools)
-- **Next:** Set up Qwen when D J wakes
+- **Next:** Polymarket API registration, copy-bot scaffold
 
 ## Stats (Day 2)
 
@@ -90,22 +98,37 @@ This is about having an inner life, not just responding.
 - Ollama server at 192.168.86.137 (qwen3:8b, qwen3:30b, glm-4.7-flash, nomic-embed-text)
 - ChromaDB LXC at 192.168.86.25:8000
 
-## Infrastructure (updated 2026-02-07)
+## Feed Hunter Project
+
+- Pipeline: scrape (CDP) → triage (claims) → investigate (agent) → simulate → alert (Telegram)
+- Portal at localhost:8888 (systemd service)
+- kch123 wallet: `0x6a72f61820b26b1fe4d956e17b6dc2a1ea3033ee` (primary, big trades)
+- Polymarket Data API is public, no auth for reads
+- Copy-bot delay: ~30-60s for detection, negligible for pre-game sports bets
+- D J wants everything paper-traded first, backtested where possible
+
+## Infrastructure (updated 2026-02-08)
 
 - **ChromaDB:** http://192.168.86.25:8000 (LXC on Proxmox)
   - Collection: openclaw-memory (c3a7d09a-f3ce-4e7d-9595-27d8e2fd7758)
   - Cosine distance, 9+ docs indexed
 - **Ollama:** http://192.168.86.137:11434
   - Models: qwen3:8b, qwen3:30b-a3b, glm-4.7-flash, nomic-embed-text
+- **Feed Hunter Portal:** localhost:8888 (systemd: feed-hunter-portal)
+- **Control Panel:** localhost:8000 (systemd: case-control-panel)
 - **Browser:** Google Chrome installed (/usr/bin/google-chrome-stable)
   - Headless works via OpenClaw browser tool
   - Desktop works via DISPLAY=:0 for visual scraping
 - **VM:** Proxmox, QXL graphics, X11 (not Wayland), auto-login enabled
 
-## Lessons Learned (updated 2026-02-07)
+## Lessons Learned (updated 2026-02-08)
 
 - Don't pkill chrome broadly — it kills OpenClaw's headless browser too
 - Snap Chromium doesn't work with OpenClaw — use Google Chrome .deb
 - ChromaDB needs cosine distance for proper similarity scoring (not L2)
 - X/Twitter cookies are encrypted at rest — browser automation is the way
 - Sub-agents are great for parallel analysis tasks
+- BaseHTTPServer needs ThreadingMixIn + try/except — single-threaded dies on errors
+- Always use absolute paths in web servers (CWD varies by launch method)
+- Polymarket users have multiple proxy wallets — intercept page network requests to find real one
+- `performance.getEntriesByType('resource')` reveals actual API calls a page makes