social-science-skills

Pass

Audited by Gen Agent Trust Hub on Apr 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The collection provides instructional content and code examples for social science research methods, including survey design and social network analysis.
  • [SAFE]: API integration in the ipums-microdata-api skill follows best practices by using environment variables for credential management and targeting legitimate, well-known academic endpoints at api.ipums.org.
  • [SAFE]: Python dependencies referenced in the skills, such as pandas, scipy, and networkx, are standard and well-maintained libraries within the scientific computing ecosystem.
  • [PROMPT_INJECTION]: Evaluation of indirect prompt injection surfaces indicates that while the skills provide code to ingest external data (e.g., loading CSV files for analysis in psychology-research-guide/SKILL.md), the scripts do not utilize high-risk capabilities such as arbitrary command execution or dynamic code evaluation.
  • Ingestion points: Data loading via pandas in psychology-research-guide/SKILL.md and network-analysis-guide/SKILL.md.
  • Boundary markers: Not implemented in the provided code templates.
  • Capability inventory: Data processing and statistical computation using established libraries. No subprocess, exec, or eval usage was detected.
  • Sanitization: Not present in the provided snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 13, 2026, 04:35 AM
Security Audit — agent-trust-hub — social-science-skills