academic-paper

Pass

Audited by Gen Agent Trust Hub on May 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, prompt injections, or exfiltration attempts were detected. The skill is thoroughly documented and focuses exclusively on its stated purpose of academic writing.
  • [COMMAND_EXECUTION]: The skill uses established tools like Pandoc, Tectonic, and XeLaTeX for document formatting and PDF generation. These commands are invoked in a controlled manner by the formatter agent to produce final manuscripts, which is consistent with the skill's purpose.
  • [REMOTE_CODE_EXECUTION]: The visualization agent generates Python (matplotlib/seaborn) and R (ggplot2) code to create publication-quality charts. This is the intended functionality of the agent to assist researchers with data representation.
  • [DATA_EXPOSURE]: The skill processes user-provided research materials and datasets. It includes specific protocols like 'Knowledge Isolation' to ensure that the AI focuses on these materials rather than internal training data, which serves as a security best practice for accuracy and integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
May 21, 2026, 05:53 PM
Security Audit — agent-trust-hub — academic-paper