Code-addressed node IDs and remove JSON index files

- Compiler now generates SHA3-256 hashes for node IDs
- Each hash includes type, config, and input hashes (Merkle tree)
- Same plan = same hashes = automatic cache reuse

Cache changes:
- Remove index.json - filesystem IS the index
- Files at {cache_dir}/{hash}/output.* are source of truth
- Per-node metadata.json for optional stats (not an index)

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

app/services/run_service.py

@@ -522,10 +522,11 @@ class RunService:
         if not run:
             return None
 
-        plan_cache_id = run.get("plan_cache_id")
-        if plan_cache_id:
+        # Check plan_cid (stored in database) or plan_cache_id (legacy)
+        plan_cid = run.get("plan_cid") or run.get("plan_cache_id")
+        if plan_cid:
             # Get plan from cache by content hash
-            plan_path = self.cache.get_by_cid(plan_cache_id)
+            plan_path = self.cache.get_by_cid(plan_cid)
             if plan_path and plan_path.exists():
                 with open(plan_path) as f:
                     content = f.read()