academic-research
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill suite contains an attack surface for indirect prompt injection, which is a standard risk for agents that ingest untrusted external data.
- Ingestion points: The
paper-reading/SKILL.mdskill processes content from local PDF paths and fetches data from arXiv URLs. Theprofessor-fit-analyser/SKILL.mdskill fetches data from Google Scholar, personal lab pages, and various bibliographic databases (e.g., PubMed, DBLP). - Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent interpolates fetched content into its context.
- Capability inventory: The skills leverage the agent's core capabilities for file system access (reading PDFs) and web browsing (searching and fetching publication metadata).
- Sanitization: Absent. The skill instructions do not require the agent to sanitize or filter the retrieved external content before analysis.
Audit Metadata