handoff
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill compacts the current conversation into a document intended for another agent to process. This creates a multi-step chain where untrusted input from a user or external tool in the current session could influence the behavior of the next agent picking up the handoff.
- Ingestion points: Ingests the "current conversation" content from the agent's context (SKILL.md).
- Boundary markers: The instructions do not specify any delimiters or XML tags to wrap the summarized content to prevent the next agent from interpreting summary data as instructions.
- Capability inventory: The skill has the capability to write to the file system ("Save to the temporary directory of the user's OS").
- Sanitization: The skill includes a specific instruction to "Redact any sensitive information, such as API keys, passwords, or personally identifiable information," which addresses data exposure but not the risk of instruction injection.
Audit Metadata