backend-craft

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill enforces production-grade security standards, including the use of Helmet for security headers, Zod for schema validation, and Pino for structured logging. It explicitly prohibits hardcoded credentials and unsafe CORS configurations.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user prompts to determine architecture, creating a surface for indirect instructions. * Ingestion points: User requirements containing keywords like 'backend' or 'auth' in SKILL.md. * Boundary markers: No explicit delimiters are used to isolate user input from the skill's logic. * Capability inventory: The agent can generate full backend codebases and database schemas in SKILL.md. * Sanitization: The skill focuses on sanitizing outputs via Zod rather than the input prompt itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 11:36 AM
Security Audit — agent-trust-hub — backend-craft