parse-cv

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill's primary function is to convert untrusted user files (PDF, DOCX, RTF) into plain text for LLM consumption. Malicious instructions embedded within these documents could be interpreted by the downstream agent as authoritative commands.
  • Ingestion points: File paths passed to the script, specifically located in the $JHT_HOME/profile/sources/ directory.
  • Boundary markers: The script outputs raw extracted text directly to stdout without using delimiters (e.g., XML tags or clear separators) or providing explicit instructions to the agent to disregard any commands found within the text.
  • Capability inventory: The skill leverages tool calls to pdftotext, pandoc, and cat to read file contents.
  • Sanitization: No sanitization, escaping, or filtering is applied to the extracted text before it is returned to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 12:59 AM
Security Audit — agent-trust-hub — parse-cv