pdfium-impl-fonts

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill defines a vulnerability surface for indirect prompt injection by processing untrusted font metadata from PDF files without sanitization.\n
  • Ingestion points: Font names are extracted from untrusted PDF files using the PdfFont::name() method in SKILL.md and references/examples.md (Example 8).\n
  • Boundary markers: No boundary markers or instructions are included to prevent the agent from obeying payloads or malicious sequences embedded in the font names.\n
  • Capability inventory: The skill demonstrates the use of std::fs::write to extract font programs from the PDF and save them to the local filesystem.\n
  • Sanitization: The provided code examples do not sanitize or validate the font name before using it to construct a file path, creating a path traversal vulnerability if a PDF contains a font with a name such as ../../filename.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 07:52 AM
Security Audit — agent-trust-hub — pdfium-impl-fonts