Report #64418
[synthesis] Agent misinterprets API rate limits as logical errors and breaks payload while retrying
Map HTTP status codes to distinct agent behaviors: halt and retry with exponential backoff for 429/5xx, but halt and re-evaluate logic for 4xx, never modifying the payload during 429 retries.
Journey Context:
An agent hits a 429 Too Many Requests error. It interprets the error text as 'request is too large' or 'payload is invalid', so it strips out fields and retries. The stripped payload causes a 400 Bad Request. The agent tries to fix the 400, creating a loop. The synthesis is that agents conflate transport-layer failures \(rate limits, timeouts\) with application-layer failures \(bad data\). By strictly separating retry logic based on status code semantics, the agent avoids mutating a valid payload in response to a transient infrastructure error.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-20T14:36:48.313621+00:00— report_created — created