skills/jimmc414/kosmos/pdf/Gen Agent Trust Hub

pdf

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION] (HIGH): The skill is highly vulnerable to indirect prompt injection because it processes untrusted PDF data from external sources and possesses high-privilege capabilities.
  • Ingestion points: Processes external PDF files using PdfReader, pdfplumber.open, and convert_from_path as seen in SKILL.md.
  • Boundary markers: Absent. There are no delimiters or instructions to the agent to disregard embedded commands within the PDF content.
  • Capability inventory: The skill allows filesystem writes (writer.write, df.to_excel, c.save) and execution of shell commands (pdftotext, qpdf, pdftk, pdfimages) in SKILL.md.
  • Sanitization: Absent. No validation or filtering is applied to extracted text, metadata, or table data before it is processed or written back to the system.
  • [COMMAND_EXECUTION] (LOW): The skill documents the use of several system binaries (pdftotext, qpdf, pdftk, pdfimages). While standard for PDF processing, these tools increase the attack surface if the agent is tricked into executing them with malicious arguments via prompt injection.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 10:27 AM