data-analysis

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill generates HTML reports that load the 'Chart.js' library from 'jsDelivr' and fonts from 'Google Fonts'. These are well-known services used for visualization and typography.
  • [PROMPT_INJECTION]: The skill processes user-uploaded data, which presents a surface for indirect prompt injection or cross-site scripting in the generated report.
  • Ingestion points: User-provided CSV and Excel files defined in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions to disregard embedded commands in the data are provided to the model.
  • Capability inventory: The agent uses 'pandas' to read files and generates a local HTML report file.
  • Sanitization: The instructions do not explicitly require the model to sanitize data values before including them in the HTML template's JavaScript block.
  • [SAFE]: The core functionality of the skill follows established data analysis workflows and utilizes common Python packages without any evidence of malicious intent or unauthorized actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 01:54 AM
Security Audit — agent-trust-hub — data-analysis