Skills
▥AGENT LAB: SKILLS
skills/daymade/claude-code-skills/markdown-tools/Gen Agent Trust Hub

markdown-tools

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (LOW): Indirect Prompt Injection Surface. The skill converts untrusted documents (PDF, DOCX, PPTX, XLSX) into Markdown for LLM processing. \n
  • Ingestion points: scripts/convert.py (referenced) and scripts/extract_pdf_images.py ingest user-supplied documents. \n
  • Boundary markers: Absent. The documentation and scripts do not describe or implement markers to prevent the agent from following instructions embedded in the documents. \n
  • Capability inventory: The skill allows file system write operations (extracted images) and execution of external tools via uv run and subprocess calls. \n
  • Sanitization: Absent. Data is extracted and formatted as Markdown without sanitization of the underlying text content.\n- EXTERNAL_DOWNLOADS (SAFE): The skill references installation of well-known and trusted tools such as markitdown (Microsoft), pymupdf4llm, and pandoc. These are standard tools for document processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:00 PM