handoff
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection vulnerability surface.
- Ingestion points: Current conversation history and user-supplied arguments (SKILL.md).
- Boundary markers: None; the skill does not use delimiters to distinguish between conversation data and agent instructions.
- Capability inventory: File system write access (targeted at the OS temporary directory).
- Sanitization: The skill includes instructions to redact credentials and personally identifiable information (PII), but lacks measures to sanitize potential instructions embedded in the summarized data.
- [DATA_EXFILTRATION]: The skill instructs the agent to save the handoff document in the operating system's temporary directory rather than the project workspace. Storing conversation summaries in a shared system location may lead to unintended data exposure depending on system permissions.
- [SAFE]: The skill demonstrates a safety-conscious design by explicitly requiring the agent to redact sensitive information like API keys, passwords, and PII before writing the file.
Audit Metadata