pdf

Fail

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (HIGH): The skill is susceptible to Indirect Prompt Injection (Category 8) because it ingests untrusted PDF data and has write/execute capabilities. * Ingestion points: PdfReader and pdfplumber.open in SKILL.md. * Boundary markers: Absent. * Capability inventory: File-writing (writer.write, canvas.save) and system command execution (qpdf, pdftotext). * Sanitization: Missing for external content.
  • [EXTERNAL_DOWNLOADS] (LOW): References standard libraries pypdf, pdfplumber, reportlab, pandas, and pytesseract. These are considered low risk given the trusted source (anthropics/skills) per the [TRUST-SCOPE-RULE].
  • [COMMAND_EXECUTION] (MEDIUM): Employs CLI tools like pdftotext and qpdf. While standard for PDF processing, these execute on potentially malicious untrusted input files.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 15, 2026, 10:19 PM