what
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No security risks were identified. The skill acts as a formatting and summarization prompt for existing conversation history.
- [INDIRECT_PROMPT_INJECTION]: The skill processes conversation history and tool results, which are potentially untrusted ingestion points. However, the instructions explicitly prohibit executing new tools and require a strict plain-text output format, which eliminates the possibility of downstream capability abuse. 1. Ingestion points: Reads the last message and previous tool results from the conversation history (specified in SKILL.md). 2. Boundary markers: The skill defines a strict output contract but does not implement specific input delimiters. 3. Capability inventory: None. Instructions explicitly forbid running new tools to reconstruct facts. 4. Sanitization: No explicit content filtering or escaping is implemented.
Audit Metadata