pdf-skill

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill provides instructions for parsing and extracting data from PDF files, which is a common vector for indirect prompt injection.
  • Ingestion points: PDF parsing and text extraction tasks described in SKILL.md (e.g., using pdf-parse, PyPDF2).
  • Boundary markers: No specific delimiters or safety instructions are provided to separate extracted content from agent instructions.
  • Capability inventory: The skill involves file system operations (saving PDFs) and potential network activity (via Puppeteer/Playwright for HTML conversion).
  • Sanitization: There is no guidance on sanitizing or validating the content extracted from PDFs before it is used in subsequent reasoning steps.
  • [External Downloads] (SAFE): The skill references standard, reputable libraries (e.g., pdfkit, puppeteer, pdf-lib) from public registries. These references are for legitimate functionality and do not involve suspicious remote script execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 12:18 PM