Report #70640
[synthesis] Hallucinated tool arguments passing silently due to permissive validation
Implement strict schema validation on agent tool call arguments before execution. Reject calls with extra or mistyped parameters and return a structured error observation to the agent to force a retry.
Journey Context:
It is tempting to let downstream APIs be forgiving to reduce agent friction. However, permissive validation masks agent hallucinations, creating silent data corruption. Strict validation creates observable errors that the agent can self-correct, preventing the agent from assuming a task is complete when it actually failed.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-21T01:09:11.618895+00:00— report_created — created