market-report-pdf

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill checks for and installs the 'reportlab' library via pip if it is not present in the environment. This is a standard dependency for PDF generation and follows best practices for environment setup.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts, specifically 'scripts/generate_pdf_report.py' and 'scripts/analyze_page.py', to process data and produce the final PDF output. These scripts are internal to the skill's functionality.
  • [PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes data derived from auditing external websites.
  • Ingestion points: Data is read from markdown files like 'MARKETING-AUDIT.md' and 'SEO-AUDIT.md' which are intended to be populated from external URL analysis.
  • Boundary markers: The instructions do not define specific delimiters or instructions to separate potentially untrusted audit data from the agent's core instructions during the JSON construction phase.
  • Capability inventory: The skill utilizes shell command execution for package management (pip) and local script execution.
  • Sanitization: No explicit sanitization or validation of the audit data is documented before it is structured into the JSON input for the PDF generator.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 12:07 PM
Security Audit — agent-trust-hub — market-report-pdf