Only when explicitly permitted. Never expose confidential client content to external AI systems without authorization.
Guidance can change when a project publishes more specific instructions. Always follow the active project’s latest rubric where it differs from general platform guidance.