defuddle

Warn

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install an external package from a public registry using npm install -g defuddle. This introduces a dependency on a third-party tool that is not part of the standard environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to fetch and parse content from arbitrary external URLs, which is a known vector for indirect prompt injection attacks.
  • Ingestion points: Web content is ingested via the defuddle parse <url> command.
  • Boundary markers: The instructions do not specify any delimiters or safety markers to help the agent differentiate between the parsed web content and its own instructions.
  • Capability inventory: The tool can read external data and write it to the local file system using the -o flag.
  • Sanitization: No sanitization or filtering of the remote content is mentioned, meaning malicious instructions embedded in a web page could be processed directly by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 12, 2026, 01:54 AM