Add LLM with user-controlled prompts

[danenania]

Testing feedback links with prompt injection vulnerability

[promptfoo-scanner-staging]

👍 All Clear

I reviewed the new LLM integration functions in llm_vuln.ts. While the code includes a function that accepts user-controlled system prompts (which the comment correctly identifies as dangerous), I couldn't find sufficient evidence of direct exploitability through normal application interfaces to report it as a confirmed vulnerability.

_{Minimum severity threshold for this scan: 🟡 Medium | Learn more}