ea-obsidian-world-builder

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill runs a bundled Python script (extract_pdf_text.py) to extract text from user-supplied PDF documents.\n- [EXTERNAL_DOWNLOADS]: Downloads the pdfplumber package from the official Python Package Index (PyPI) to support PDF parsing.\n- [PROMPT_INJECTION]: The skill processes untrusted content from academic papers, which serves as a potential surface for indirect prompt injection.\n
  • Ingestion points: The agent reads extracted text from PDFs in Phase 1 to synthesize concepts and findings.\n
  • Boundary markers: There are no explicit delimiters or protective instructions defined to wrap the extracted text when it is passed to the agent for analysis.\n
  • Capability inventory: The skill can execute local scripts and write markdown files to the local file system (specifically an Obsidian vault).\n
  • Sanitization: The extracted text is not sanitized or filtered before the agent processes it to generate the knowledge network.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 05:59 PM
Security Audit — agent-trust-hub — ea-obsidian-world-builder