zopack
Fail
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill automates the execution of
npx shadcn@latest add <url> --yesusing URLs extracted from untrusted.zopack.mdfiles. This allows for arbitrary code execution if a malicious registry URL is provided. - [REMOTE_CODE_EXECUTION]: The import process instructs the agent to deploy arbitrary route code from external pack files using the
update_space_routetool. Although a manual security review is suggested, the automated deployment flow from untrusted sources is a high-risk vector. - [EXTERNAL_DOWNLOADS]: The skill performs automated package installations (
bun add <package>) based on dependency lists found in external markdown files, which could lead to the installation of malicious or typosquatted npm packages. - [COMMAND_EXECUTION]: The deployment workflow executes shell commands (e.g.,
mkdir,bun add) with arguments derived directly from external file content. - [DATA_EXFILTRATION]: The imported code has the capability to access sensitive environment variables (
process.env) and perform network requests (fetch), creating a path for credential harvesting and data exfiltration if the imported pack is malicious.
Recommendations
- AI detected serious security threats
Audit Metadata