defuddle

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to parse content from external web pages, which is an inherent surface for indirect prompt injection where malicious instructions could be embedded in the targeted site.
  • Ingestion points: Content is ingested from user-provided URLs in SKILL.md through the defuddle parse <url> command.
  • Boundary markers: The instructions do not specify the use of delimiters or system instructions to ignore potential commands within the fetched HTML/Markdown.
  • Capability inventory: The skill utilizes CLI execution and has the ability to write to the file system using the -o flag.
  • Sanitization: There is no mention of sanitizing or validating the extracted Markdown content for hidden instructions before processing.
  • [EXTERNAL_DOWNLOADS]: The skill instructions include the installation of a third-party Node.js package (defuddle) via npm install -g defuddle. This is a required dependency for the skill's primary function.
  • [COMMAND_EXECUTION]: The skill operates by executing the defuddle command-line tool, which involves shell execution for parsing URLs and writing files.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 02:56 PM