handoff
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill compacts untrusted conversation history into a handoff document for use by another agent, creating a multi-step indirect prompt injection surface.\n
- Ingestion points: The skill ingests the entire "current conversation" (SKILL.md).\n
- Boundary markers: Absent. There are no instructions to use delimiters or ignore instructions embedded in the conversation.\n
- Capability inventory: The agent is instructed to write a file to the system's temporary directory, which is a file-system write capability (SKILL.md).\n
- Sanitization: The skill includes instructions to redact PII and secrets, but does not provide sanitization or escaping mechanisms for potentially malicious instructional text within the processed conversation.
Audit Metadata