← Automation Failure Diagnostic

Sample automation failure diagnostic

Sample · self-directed · non-production · synthetic data onlyThis sanitized sample reviews the public lead-triage demonstration in this repository. It is not client work, a production deployment, or evidence of production results. Observations come from source inspection and locally executed tests using synthetic fixtures.

Executive summary

The sample workflow correctly rejects a lead with no email, returns a visible correction route, and produces matching results in the companion JavaScript and the JavaScript embedded in its n8n Code node for the tested fixture.

The most important production-readiness issue is response semantics: the workflow is configured to return HTTP 200 even when its body says ok: false. A sender or monitor relying only on HTTP status could record invalid input as accepted. Before attaching a CRM or outbound action, the path would also need explicit input ranges, duplicate protection, and verification in the intended n8n runtime.

Bounded scope

Synthetic webhook body → normalize and validate → score and route → JSON response

Reviewed: the workflow JSON, companion JavaScript, behavioral tests, parity tests, and one local test run using synthetic fixtures. Excluded: n8n import or execution, credentials, live traffic, external writes, load or security testing, repair implementation, and business-performance claims.

Acceptance test and observed result

Given the synthetic input below, the implementation must return ok: false, route to needs-correction, include an actionable email error, and match the embedded Code-node result.

{
  "name": "No Email"
}

Observed synthetic result:

{
  "ok": false,
  "route": "needs-correction",
  "errors": ["a valid email is required"]
}

Local test summary: 4 passed, 0 failed. This proves only the tested JavaScript behavior; it does not emulate n8n or demonstrate a production webhook.

Ranked findings

Priority	Finding	Recommended control
High	Rejected input is configured to receive HTTP 200.	Agree and test a 4xx response or a documented body-aware caller contract.
High before writes	No idempotency or duplicate control exists.	Define a stable event key and test the same accepted event twice.
Medium	Numeric fields are coerced but not range-validated.	Reject or normalize values outside agreed integer ranges.
Medium	The routing threshold boundary is not directly tested.	Add fixtures immediately below and at the threshold.
Release blocker	Behavior in the target n8n runtime is unverified.	Import and execute valid and invalid fixtures in the intended n8n version.

The HTTP-status finding is directly observable in the artifact. The others are risks inferred through source review, not claims that a production incident occurred.

Smallest credible repair

Starting at USD 250: agree the invalid-input contract; add appropriate status behavior and numeric range validation; test valid input, invalid email, and the routing boundary; then deliver the updated artifact, test evidence, known gaps, and handoff notes. Final scope and price are confirmed in writing first.

Apply this format to one failing path

The fixed diagnostic is USD 125, credited in full toward an agreed USD 250+ repair on the same path.

Request a written scope →

Do not send credentials or customer data. No production action is performed during the diagnostic.