defuddle
Warn
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install a global Node.js package 'defuddle' (npm install -g defuddle). This package originates from a source not recognized as a trusted vendor, posing a supply chain risk.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It ingests data from external URLs (defuddle parse ) and converts it into markdown for agent consumption. * Ingestion points: The argument in the defuddle command defined in SKILL.md. * Boundary markers: The instructions lack delimiters or warnings to treat fetched content as untrusted data. * Capability inventory: The skill uses the defuddle CLI and has the ability to write output to the filesystem. * Sanitization: There is no mention of sanitizing or filtering the content retrieved from external URLs.
- [COMMAND_EXECUTION]: The skill executes a CLI tool where user-controlled input (URLs) is passed as a command-line argument. If the URL is not properly sanitized, it could potentially be exploited for command injection depending on the environment's shell handling.
Audit Metadata