url-reader
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes curl commands to interact with an external service at defuddle.md.
- [DATA_EXFILTRATION]: The skill transmits user-supplied URLs to a third-party extraction service. The instructions include a explicit warning to avoid sending private, login-gated, or sensitive URLs, which serves as a mitigation for data exposure.
- [EXTERNAL_DOWNLOADS]: The skill downloads content from remote sources via a proxy service, creating a dependency on the integrity of the external content provider.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from the webpages it summarizes. 1. Ingestion points: Webpage content retrieved via curl from defuddle.md. 2. Boundary markers: The skill lacks explicit markers or delimiters to encapsulate the untrusted external content. 3. Capability inventory: The skill utilizes network access via curl. 4. Sanitization: No content sanitization or instruction-filtering is described in the skill instructions.
Audit Metadata