UNPKG

ubon

Version:

Security scanner for AI-generated apps (Cursor, Lovable, Windsurf, v0). Catches hardcoded secrets, prompt injection, hallucinated imports, Server Actions / Edge runtime mistakes, and the vibe-coded vulnerabilities traditional linters miss.

17 lines 695 B
import { Rule } from '../types'; /** * AI001: LLM provider API key hardcoded. * * Catches OpenAI, Anthropic, Google AI Studio / Gemini, Cohere, Mistral, * Groq, Together, Perplexity and Replicate keys when they appear inline * in source code or `.env` files (which is fine if `.env` is gitignored, * but ENV001/ENV002 cover that orthogonally). * * Why a dedicated rule (vs SEC001/SEC014): AI providers ship with their own * key prefixes, fast key-rotation flows, and free-tier abuse vectors that * deserve a focused fingerprint. A leaked AI key burns budget within * minutes once it hits the wild. */ declare const rule: Rule; export default rule; //# sourceMappingURL=AI001.d.ts.map