find-ebook
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions are specifically designed to assist users in finding legitimate ebook resources. It mandates the use of official channels such as author/publisher websites, licensed digital bookstores (e.g., Kindle, Apple Books, WeChat Reading), and established library catalogs (e.g., WorldCat). Furthermore, it includes strict directives to exclude unauthorized download sites, pirated content, and methods for bypassing access restrictions.\n- [INDIRECT_PROMPT_INJECTION]: The skill operates by retrieving and analyzing information from external websites and search results, which constitutes a functional attack surface for indirect prompt injection. \n
- Ingestion points: The agent processes external data from search engine summaries and content extracted from candidate resource pages across various domains.\n
- Boundary markers: The skill incorporates a mandatory target restatement step and follows a strict markdown output schema to maintain context integrity.\n
- Capability inventory: The workflow involves web searching and browsing to verify book details (ISBN, authorship) and access terms.\n
- Sanitization: The skill mitigates risks through a mandatory verification phase (Step 4) that requires the agent to validate the authenticity and authorization status of each source, ensuring that only verified content is presented to the user.
Audit Metadata