feat: L1 server URL sent with publish request (many-to-many support)

- Add l1_server field to RecordRunRequest
- L2 fetches run data from the specified L1 URL instead of hardcoded config
- Store l1_server in provenance and metadata
- Remove ARTDAG_L1 config requirement from L2
- Update docker-stack.yml comments

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

.env.example

@@ -6,8 +6,7 @@ ARTDAG_DOMAIN=artdag.rose-ash.com
 # Default username (for actor endpoints)
 ARTDAG_USER=giles
 
-# L1 server URL (for fetching run data)
-ARTDAG_L1=https://l1.artdag.rose-ash.com
-
 # JWT secret for token signing (generate with: openssl rand -hex 32)
 JWT_SECRET=your-secret-here-generate-with-openssl-rand-hex-32
+
+# Note: ARTDAG_L1 is no longer needed - L1 server URL is sent with each request