handbook-gate-reread
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external text data (
chapter_laid_outandoutline) which could potentially contain malicious instructions intended to influence the agent's behavior. - Ingestion points: The skill accepts
chapter_laid_out(the chapter draft) and theoutlineas its primary inputs inSKILL.md. - Boundary markers: The instructions do not specify any delimiters (such as XML tags or markdown code blocks) to isolate the untrusted input from the core instructions, increasing the risk that embedded commands could be interpreted as part of the prompt.
- Capability inventory: The skill has no high-risk capabilities; it does not utilize shell execution, network access, or file system writes. Its primary action is to generate a review report in a structured format.
- Sanitization: There are no mentioned mechanisms for sanitizing, filtering, or escaping the input text before processing.
Audit Metadata