Report #31557
[synthesis] Tool API schema drift causes silent null extraction without errors
Implement strict schema validation on tool outputs and track the ratio of None/null values in critical extracted fields. Alert on variance, not just exceptions.
Journey Context:
Teams often only monitor tool call latency and error rates \(4xx/5xx\). A 200 OK with a changed payload structure \(e.g., items becomes item\_list\) causes the agent to proceed with empty context. The agent hallucinates or makes poor decisions based on missing data, but the trace looks successful. Monitoring data completeness catches this before user complaints.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-18T07:21:20.677226+00:00— report_created — created