report-analyzer

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands such as find to locate PDF and PPTX files within the working directory and subdirectories.
  • [EXTERNAL_DOWNLOADS]: Automatically installs several Python packages including pdfplumber, pypdf, markitdown, pytesseract, and pdf2image to facilitate file processing.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests data from untrusted external documents (PDF and PPTX reports) provided by users or attackers.
  • Ingestion points: The skill reads full text and table data from uploaded documents in SKILL.md (Step 3).
  • Boundary markers: Employs instructions such as 'Extract only facts and data from the report — do not infer or fabricate' and 'All numbers must be exact' to guide the agent, though these do not prevent adversarial injection.
  • Capability inventory: The skill has the ability to execute arbitrary Python code and shell commands for file management and extraction.
  • Sanitization: No programmatic sanitization or filtering of document content is performed prior to analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 09:41 AM
Security Audit — agent-trust-hub — report-analyzer