How I found and fixed real schema errors in n8n workflows

This one shows the actual before/after evidence, because that's the part that's usually missing.

The problem

n8n workflow JSON files accumulate schema drift over time — nodes get added,
connections get edited by hand, and small structural errors creep in that don't
show up until a workflow fails in production. I wanted a tool that could
actually find and fix these, with every change logged, not a black box.

What I built

A workflow analyzer and auto-fix pipeline that:

• Parses an n8n workflow's JSON structure and checks it against n8n's own schema expectations
• Flags structural issues (broken connections, malformed node parameters, schema mismatches)
• Applies fixes where it's confident, and logs exactly what changed

The real numbers

On the test workflow I ran this against:

• 11 findings before the fix pass
• 5 schema edits applied
• 6 findings remained after (documented, not hidden — some issues need a human decision, not an automated fix)
• 16 nodes retained unchanged

I'm not claiming 100% automated fixing — I'm showing the actual before/after
with the limitations visible, because that's more useful than a demo that only
shows the happy path.

Full writeup and the live evidence page: https://jamilfayad.com/projects/n8n-workflow-autofix.html

Happy to answer questions about the schema-validation approach or n8n's JSON
structure in the comments.