Reset stale agent runs before startup to prevent stack overflow

Paperclip's startup recovery does recursive DFS on the issue graph.
Deeply nested stalled-task chains cause RangeError on restart.
Marking orphaned running runs as failed before launch prevents the
recovery code from firing entirely.

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

start.sh

@@ -212,6 +212,9 @@ cleanup() {
 }
 trap cleanup SIGTERM SIGINT
 
+# Reset stale running agent runs — prevents stack overflow in startup recovery
+su - postgres -c "psql paperclip -c \"UPDATE agent_run SET status='failed' WHERE status='running';\"" >/dev/null 2>&1 || true
+
 # ── Launch Paperclip ──────────────────────────────────────────────────────────
 echo "Starting Paperclip..."
 node --import ./server/node_modules/tsx/dist/loader.mjs server/dist/index.js &