academic-research

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection by processing untrusted data from an external academic API.
  • Ingestion points: Paper metadata, including titles, abstracts, and TLDR summaries, are ingested from api.semanticscholar.org via the search_papers.py script.
  • Boundary markers: While output is provided in structured JSON or BibTeX formats, the agent instructions lack boundary markers or explicit directions to disregard instructions found within the scholarly content.
  • Capability inventory: The skill possesses network access to a well-known service and the ability to output results to the console; it does not have the capability to write to the file system or execute arbitrary shell commands.
  • Sanitization: There is no evidence of sanitization or filtering of the retrieved text to remove potential injection strings before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 09:05 PM