ptns
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill generates a handoff prompt for a new session based on the current session's history and state. This introduces a surface for indirect prompt injection, as malicious content or instructions present in the current conversation could be carried over and executed by the next session.
- Ingestion points: The skill gathers state from the current conversation history and environment as defined in SKILL.md.
- Boundary markers: There are no instructions to wrap the generated handoff content in delimiters or to include warnings for the receiving session to treat the data as untrusted.
- Capability inventory: The skill utilizes the ListAgents and SendMessage tools to facilitate the transfer of state between active sessions (SKILL.md).
- Sanitization: The skill proactively mitigates data exposure by instructing the agent to redact credentials (API keys, passwords, personal data) using placeholders. However, it lacks sanitization or escaping mechanisms for natural language instructions that might be contained within the history being summarized.
Audit Metadata