pdf-extract-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documents the use of the pdftotext utility and provides instructions for its installation via the system package manager (dnf).
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the poppler-utils package from the system's official package repository using dnf.
  • [DATA_EXFILTRATION]: The skill explicitly instructs the agent to avoid piping output to network-transmitting commands, addressing risks associated with unauthorized data transmission.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external PDF files, which represents an indirect prompt injection surface. Ingestion points: PDF documents processed through pdftotext (SKILL.md). Boundary markers: The instructions recommend returning extracted text directly to the user for review but do not define technical delimiters. Capability inventory: Shell access for command execution (SKILL.md). Sanitization: No programmatic sanitization is defined for the extracted document content.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:32 PM
Security Audit — agent-trust-hub — pdf-extract-hardened