defuddle

Warn

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install a third-party package defuddle globally using npm install -g defuddle. This introduces a dependency from the public npm registry.
  • [COMMAND_EXECUTION]: The skill executes the defuddle CLI tool on the local system, passing user-provided URLs as arguments.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from external websites. Ingestion points: Untrusted web content enters the agent's context through the defuddle parse <url> command. Boundary markers: No specific delimiters or "ignore instructions" warnings are provided to separate fetched web content from the agent's system prompt. Capability inventory: The skill itself uses CLI execution, and while it doesn't directly pipe tool output to dangerous functions within the skill, the agent reading the output possesses general capabilities. Sanitization: No sanitization, validation, or filtering of the extracted markdown content is performed before it is processed by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 21, 2026, 09:41 AM
Security Audit — agent-trust-hub — defuddle