landing-page
Warn
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The script 'generate-project.js' utilizes 'child_process.execSync' to execute other Node.js generator scripts. It passes the user-supplied 'projectName' directly into a shell command template string. This creates a command injection vulnerability if a user provides a project name containing shell metacharacters (e.g., quotes, semicolons, or pipes).
- [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection attack surface. In 'SKILL.md' (Phase 1, Q0), the agent is instructed to ask the user for a file path to 'Landing Page Worksheets' and read the contents using a Read tool. This external data is then used to guide the entire generation process. A malicious worksheet could contain instructions to override agent behavior or inject malicious scripts (XSS) into the generated landing page code.
- [DATA_EXFILTRATION]: Multiple scripts, including 'init-project.js' and 'generate-project.js', use 'path.join' with the unsanitized 'projectName' argument provided by the user to determine file writing locations. This allows for path traversal attacks where a user could specify names like '../../' to cause the agent to read or write files outside the intended 'projects/' directory.
Audit Metadata