landing-page-vercel
Fail
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- COMMAND_EXECUTION (HIGH): The skill requires execution of a local Python script 'scripts/scaffold.py' which is not provided for security review. Running unvetted scripts from untrusted authors ('Ship Shit Dev') can lead to full system compromise.
- PROMPT_INJECTION (HIGH): High risk of Indirect Prompt Injection. The skill ingests untrusted user input (Product Name, Tagline, Features) and uses it to generate code (HTML/JS) and data files. Evidence: 1. Ingestion points: Phase 1 Intake in SKILL.md. 2. Boundary markers: None provided in instructions to delimit user data from code. 3. Capability inventory: Local script execution and file system writes. 4. Sanitization: No sanitization or escaping of user input is specified before interpolation into generated files.
- EXTERNAL_DOWNLOADS (LOW): Recommends global installation of the Vercel CLI via npm. Although Vercel is a legitimate service, installing global packages on the host machine increases the attack surface.
Recommendations
- AI detected serious security threats
Audit Metadata