literature-filter

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): The skill instructions focus strictly on literature screening criteria and do not contain any patterns used to override agent safety or extract system prompts.
  • [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or unauthorized network exfiltration paths were detected. The skill uses legitimate tools for PubMed and Zotero.
  • [Unverifiable Dependencies & Remote Code Execution] (SAFE): No external package installations or remote script executions (e.g., curl | bash) are present.
  • [Obfuscation] (SAFE): No Base64, zero-width characters, or other hidden encoding techniques were identified.
  • [Persistence Mechanisms] (SAFE): The skill does not attempt to modify configuration files or establish scheduled tasks.
  • [Indirect Prompt Injection] (SAFE): While the skill ingests external PubMed data, it has no high-risk capabilities like writing to files or arbitrary code execution, minimizing the risk of a data-driven exploit.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:48 PM