Intent-Code Divergence
Medium
- Confidence
- 92% confidence
- Finding
- The skill instructs the agent to 'automatically create a linked Issue (or Risk)' when an assumption is invalidated, which conflicts with the broader workflow rule to ask one question at a time when required information is missing and to avoid silent reclassification. This can cause the agent to create or transform governance records without explicit user confirmation, leading to inaccurate RAID state, unauthorized status changes, and misleading downstream RAG/status reporting.