frontend-design
Fail
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDYNAMIC_CONTEXT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is configured as a stub that directs the agent to fetch its full implementation from a third-party GitHub repository (github.com/Ilm-Alan/frontend-design). This source is not verified or associated with a trusted vendor.
- [COMMAND_EXECUTION]: The instructions include specific shell commands for the agent to run, such as checking for file existence and using
grepto parse configuration files. These commands are intended to automate the transition from a stub to a full installation. - [DYNAMIC_CONTEXT_INJECTION]: The documentation explicitly tells the agent to use the
!prefix to execute commands like! npx skills add ...directly within the conversation. This pattern facilitates the immediate execution of shell commands to install external, unverified code. - [REMOTE_CODE_EXECUTION]: The suggested installation method uses
npxto add a skill from an external repository. This allows for the download and subsequent execution of remote code within the agent's local environment without manual source code review.
Recommendations
- AI detected serious security threats
Audit Metadata