vercel
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill interacts with the official and well-known Vercel REST API (api.vercel.com) to perform administrative tasks. It uses environment variables for authentication and contains clear instructions to prevent the leakage of secret tokens or decrypted project environment variables.
- [PROMPT_INJECTION]: The skill processes untrusted external data by reading deployment build and runtime logs. While this represents a potential indirect prompt injection vector where logs could contain malicious instructions, the risk is inherent to the skill's function and is mitigated by the restricted execution environment.
- Ingestion points: Fetches log data from the Vercel API (SKILL.md)
- Boundary markers: None present for log output
- Capability inventory: Shell access (Bash) and network access (curl) (SKILL.md)
- Sanitization: None present
Audit Metadata