Direct Answer Summary
When AI coding assistants (such as Claude Code, Cursor, or OpenAI Codex) generate or refactor n8n automation pipelines, they frequently introduce subtle structural errors: dangling connection pointers, non-unique node IDs, infinite execution cycles, or exposed plaintext secrets. ELYRAON Workflow Rescue performs pure static Abstract Syntax Tree (AST) analysis on exported workflow JSON, returning an audit report of confirmed structural defects, risk classifications, and concrete repair steps without executing code or calling external networks.
1. Representative Static Checks
Workflow Rescue inspects exported n8n workflow JSON for structural and configuration findings. The published local CLI reports nine check categories; the exact hosted MCP output must be read from its current tool schema. This analysis does not execute a workflow or verify runtime behavior.
1. Broken Connection Targets
Flags edges pointing to node names or IDs absent from the nodes array.
2. Duplicate Node Identifiers
Catches duplicate IDs that cause nondeterministic routing during import.
3. Duplicate Node Names
Prevents ambiguous name-based connection resolution in n8n UI.
4. Disconnected Subgraphs
Identifies orphaned operational nodes lacking input or output connections.
5. Missing Trigger Detection
Warns when an executable workflow has no recognized webhook or schedule trigger.
6. Circular Loop Hazards
Detects cycles that cause unbounded execution loops and worker memory exhaustion.
7. Disabled Node Hazards
Surfaces disabled nodes that silently break downstream data flows.
8. Plaintext Secret Exposure
Flags unredacted tokens, passwords, and private keys in parameters.
9. Unhandled Terminal Ends
Identifies failure branches that terminate without notification or error logging.
2. Option A: Zero-Network In-Browser Inspection ($0)
When workflows contain proprietary business logic that cannot leave developer machines, run the analysis client-side:
- Open the free Workflow Rescue In-Browser Tool.
- Paste the raw n8n workflow JSON or drag-and-drop the file.
- Inspection executes 100% inside your browser's local JavaScript runtime. Zero network telemetry or server uploads occur.
- Export a structured repair checklist or sanitized JSON.
3. Option B: Automated Agent MCP Validation
When integrating directly into automated coding agent pipelines, call the live production MCP tool:
curl -s -X POST https://elyraon.com.co/agents/mcp \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Idempotency-Key: workflow-check-001" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "check_n8n_workflow",
"arguments": {
"workflow": { "nodes": [...], "connections": {...} }
}
}
}'
4. Interpreting the Audit Report
A successful HTTP 200 service call returns structured findings:
structural_pass: true/false— Indicates whether critical or high-severity syntax violations were detected.report.findings— Array of specific findings with code, severity, affected node, evidence, and exact repair advice.report.limitations— Explicit boundaries (e.g. static checks cannot verify third-party API availability or runtime execution).- The result is static diagnostic output, not a signed delivery receipt or proof of runtime success. Read the currently discovered tool schema for the exact response fields.
Next Steps for Automation Engineers
Launch the free browser tool, connect your agent via MCP, or inspect the CLI version for CI/CD pipelines.