library
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources that may be influenced by third parties, creating a surface for indirect prompt injection.\n
- Ingestion points: The agent ingests untrusted data from
CRM/calls,library evidence, and search results to populate library entities and motion ICP cells (references/task-method.md).\n - Boundary markers: The skill instructions specify the use of
protected-field boundariesandschema boundariesto restrict the scope of data operations (SKILL.md,references/task-method.md).\n - Capability inventory: The skill can execute workspace mutations, entity creation, and data search/retrieval operations within the platform environment (
SKILL.md,references/task-method.md).\n - Sanitization: Sanitization and validation mechanisms are present, including
readbackof complete saved objects before confirmation and comparing saved states with pre-images to verify only intended changes occurred (references/task-method.md).
Audit Metadata