url-to-markdown

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches converted markdown content from the established service r.jina.ai and markdown.new as a fallback.
  • [COMMAND_EXECUTION]: Executes a local Python script scripts/url_to_md.py to manage network requests and save files to the local directory using standard library modules.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its processing of untrusted external content.
  • Ingestion points: Untrusted data enters the agent context through URL fetching via urllib.request.urlopen in scripts/url_to_md.py.
  • Boundary markers: Absent. The script retrieves raw content and saves it directly to a file without applying protective delimiters or instructions to ignore embedded commands.
  • Capability inventory: The script possesses capabilities for network retrieval and local file writing via _write_output in scripts/url_to_md.py.
  • Sanitization: None. The content is saved exactly as received from the conversion service without any filtering or validation of the Markdown body.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 04:20 PM