handoff
Pass
Audited by Gen Agent Trust Hub on Jul 24, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill follows security best practices by explicitly requiring the agent to redact sensitive information, including API keys, passwords, and personally identifiable information, before any data is saved to disk.
- [PROMPT_INJECTION]: The skill summarizes conversation history, creating a surface for indirect prompt injection where untrusted user input could influence the summary or suggested actions for future agents.
- Ingestion points: The entire current conversation history is used as the source for the handoff (SKILL.md).
- Boundary markers: There are no explicit markers used to separate the conversation content from the summarization instructions.
- Capability inventory: The skill has the capability to write the summarized document to the user's temporary directory (SKILL.md).
- Sanitization: The skill includes a clear instruction to redact sensitive data (credentials and PII), mitigating potential exposure from the conversation history.
- [COMMAND_EXECUTION]: The skill instructs the agent to write output to the system's temporary directory (e.g., /tmp). This involves file system operations outside the standard project workspace, which is intended here for cross-session data persistence.
Audit Metadata