related-skill
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (LOW): The skill uses the
npx skills addcommand to download and install packages from the inference.sh registry. While this involves downloading external code from a source not explicitly on the trusted list, it is the primary function of the skill. - [COMMAND_EXECUTION] (LOW): The
allowed-toolsconfiguration grants permission to executenpx skills *commands. This provides the agent with the ability to modify its environment by adding or removing software, which is inherently risky but consistent with the skill's utility as a manager. - [REMOTE_CODE_EXECUTION] (LOW): The use of
npxinvolves fetching and executing code from a remote registry at runtime. This is categorized as low severity here because the operations are restricted to theskillspackage namespace.
Audit Metadata