backend-patterns

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface.
  • Ingestion points: The skill instructions explicitly handle data from untrusted sources including webhooks, import flows, and paste handlers (SKILL.md).
  • Boundary markers: There are no specific instructions for the agent to use delimiters or ignore embedded instructions when processing these external data sources.
  • Capability inventory: The agent is authorized to perform database writes and execute shell commands for database management (e.g., pnpm db:migrate, pnpm db:push).
  • Sanitization: No specific sanitization or validation logic is defined for the ingestion of external links beyond routing them through a helper function.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 11:02 PM
Security Audit — agent-trust-hub — backend-patterns