reviewing-ai-papers
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted technical papers and blog posts, then stores extracted insights into a persistent memory tool without sanitization.
- Ingestion points: The skill parses user-provided URLs and documents to extract AI/ML research insights as defined in SKILL.md.
- Boundary markers: There are no instructions to use delimiters or specific ignore-tags when processing external content.
- Capability inventory: The skill utilizes the 'remember' tool to store data across sessions and accesses external technical documentation.
- Sanitization: Extracted insights are interpolated directly into the memory storage function without verification or filtering of embedded instructions.
Audit Metadata