diff --git a/benchmarks/data.js b/benchmarks/data.js
index 2aec612..5d1f566 100644
--- a/benchmarks/data.js
+++ b/benchmarks/data.js
@@ -1,39 +1,8 @@
 window.BENCHMARK_DATA = {
-  "lastUpdate": 1773570603377,
+  "lastUpdate": 1773581742272,
   "repoUrl": "https://github.com/czlonkowski/n8n-mcp",
   "entries": {
     "n8n-mcp Benchmarks": [
-      {
-        "commit": {
-          "author": {
-            "email": "56956555+czlonkowski@users.noreply.github.com",
-            "name": "Romuald Członkowski",
-            "username": "czlonkowski"
-          },
-          "committer": {
-            "email": "noreply@github.com",
-            "name": "GitHub",
-            "username": "web-flow"
-          },
-          "distinct": true,
-          "id": "b92e5114634e52aee1fbb477396a21f8c229d470",
-          "message": "perf: optimize workflow tool responses for token efficiency (v2.29.0) (#479)\n\n* perf: optimize workflow tool responses for token efficiency (v2.29.0)\n\nReduce response sizes by 75-90% for 4 workflow management tools:\n\n- n8n_update_partial_workflow: Returns {id, name, active, operationsApplied}\n- n8n_create_workflow: Returns {id, name, active, nodeCount}\n- n8n_update_full_workflow: Returns {id, name, active, nodeCount}\n- n8n_delete_workflow: Returns {id, name, deleted: true}\n\nAI agents can use n8n_get_workflow with mode 'structure' if they need\nto verify the current workflow state after operations.\n\nConceived by Romuald Członkowski - https://www.aiadvisors.pl/en\n\n🤖 Generated with [Claude Code](https://claude.com/claude-code)\n\nCo-Authored-By: Claude <noreply@anthropic.com>\n\n* fix: update tests and add nodeCount to partial update response\n\n- Fix handleCreateWorkflow test to expect minimal response\n- Fix handleDeleteWorkflow test to expect minimal response\n- Add nodeCount to n8n_update_partial_workflow response for consistency\n- Update documentation and CHANGELOG\n\nConceived by Romuald Członkowski - https://www.aiadvisors.pl/en\n\n🤖 Generated with [Claude Code](https://claude.com/claude-code)\n\nCo-Authored-By: Claude <noreply@anthropic.com>\n\n* fix: update handlers-workflow-diff tests for minimal response\n\nUpdate 3 more tests that expected full workflow in response:\n- should apply diff operations successfully\n- should activate workflow after successful update\n- should deactivate workflow after successful update\n\n🤖 Generated with [Claude Code](https://claude.com/claude-code)\n\nCo-Authored-By: Claude <noreply@anthropic.com>\n\n* fix: update integration tests to use minimal response format\n\nIntegration tests now verify minimal response format and use\nclient.getWorkflow() to fetch actual workflow state for verification.\n\nConceived by Romuald Czlonkowski - www.aiadvisors.pl/en\n\n* fix: update create/update workflow integration tests for minimal response\n\nIntegration tests now verify minimal response and use client.getWorkflow()\nto fetch actual workflow state for detailed verification.\n\nConceived by Romuald Czlonkowski - www.aiadvisors.pl/en\n\n* fix: add type assertions to fix TypeScript errors in tests\n\nConceived by Romuald Czlonkowski - www.aiadvisors.pl/en\n\n---------\n\nCo-authored-by: Romuald Członkowski <romualdczlonkowski@MacBook-Pro-Romuald.local>\nCo-authored-by: Claude <noreply@anthropic.com>",
-          "timestamp": "2025-12-09T16:36:17+01:00",
-          "tree_id": "6120e3cad0bd15cbc7b0ab855182413de487bcb7",
-          "url": "https://github.com/czlonkowski/n8n-mcp/commit/b92e5114634e52aee1fbb477396a21f8c229d470"
-        },
-        "date": 1765294674834,
-        "tool": "customSmallerIsBetter",
-        "benches": [
-          {
-            "name": "sample - array sorting - small",
-            "value": 0.0136,
-            "range": "0.3096",
-            "unit": "ms",
-            "extra": "73341 ops/sec"
-          }
-        ]
-      },
       {
         "commit": {
           "author": {
@@ -1542,6 +1511,37 @@ window.BENCHMARK_DATA = {
             "extra": "73341 ops/sec"
           }
         ]
+      },
+      {
+        "commit": {
+          "author": {
+            "email": "56956555+czlonkowski@users.noreply.github.com",
+            "name": "Romuald Członkowski",
+            "username": "czlonkowski"
+          },
+          "committer": {
+            "email": "noreply@github.com",
+            "name": "GitHub",
+            "username": "web-flow"
+          },
+          "distinct": true,
+          "id": "f7a1cfe8bfe12530e012e43ca4a9de8dc943ca23",
+          "message": "fix: field normalization, AI connection validation, autofix filter (#581) (#638)\n\n- Normalize name→nodeName and id→nodeId for node-targeting operations in\n  the Zod schema transform, so LLMs using natural field names no longer\n  get \"Node not found\" errors\n- Replace hardcoded ALL_CONNECTION_TYPES with dynamic iteration so AI\n  sub-nodes (ai_outputParser, ai_document, ai_textSplitter, etc.) are\n  not flagged as disconnected during save\n- Add .catchall() to workflowConnectionSchema and extend connection\n  reference validation to cover all connection types, not just main\n- Fix filterOperationsByFixes ID-vs-name mismatch: typeversion-upgrade\n  operations now include nodeName alongside nodeId, and the filter checks\n  both fields\n\nConceived by Romuald Członkowski - https://www.aiadvisors.pl/en\n\nCo-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>",
+          "timestamp": "2026-03-15T14:32:14+01:00",
+          "tree_id": "3e8170b820822d4dc4ea7be6f3fbc5c3c099c25d",
+          "url": "https://github.com/czlonkowski/n8n-mcp/commit/f7a1cfe8bfe12530e012e43ca4a9de8dc943ca23"
+        },
+        "date": 1773581741869,
+        "tool": "customSmallerIsBetter",
+        "benches": [
+          {
+            "name": "sample - array sorting - small",
+            "value": 0.0136,
+            "range": "0.3096",
+            "unit": "ms",
+            "extra": "73341 ops/sec"
+          }
+        ]
       }
     ]
   }