reporting

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation recommends and utilizes several well-known and established Python packages, including pandas, Jinja2, Matplotlib, and WeasyPrint (specifically pinned to version 68.1). These are fetched from standard package registries during deployment (e.g., via GitHub Actions).
  • [COMMAND_EXECUTION]: The skill provides a verification script (scripts/verify.sh) intended for local use. The script performs read-only analysis of the project structure, such as verifying Jinja2 template syntax and checking for the presence of scheduled workflows. Additionally, the GitHub Actions documentation includes standard apt-get commands to install necessary system dependencies for PDF rendering.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection by ingesting external data from metrics.csv. However, this risk is appropriately mitigated by instructions to use Jinja2's autoescape=True setting and the implementation of strict report contracts that define data sources and owners.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 08:57 AM
Security Audit — agent-trust-hub — reporting