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

markitdown

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill creates a surface for indirect prompt injection by converting untrusted content (PDFs, Office documents, web pages) into Markdown intended for LLM processing.
  • Ingestion points: File paths processed in scripts/batch_convert.py and via MarkItDown.convert() calls documented in SKILL.md.
  • Boundary markers: Absent. The Markdown output is generated and written to files without specific delimiters or instructions to the LLM to ignore embedded commands within the converted text.
  • Capability inventory: The skill possesses file system write access (scripts/batch_convert.py) and network capabilities for YouTube transcript extraction and cloud-based AI integrations (Azure/OpenAI).
  • Sanitization: No sanitization or filtering of the extracted text content is identified in the provided logic.
  • External Downloads (SAFE): The skill documentation encourages the installation of the markitdown package and its dependencies from PyPI. The package is maintained by Microsoft, a trusted organization, and no suspicious installation patterns were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:14 PM