# Vouqis Verify Sample AI Change Evidence Pack

> Illustrative data only. No customer repository is shown.

## PR #184: Switch Support Agent LLM from GPT-4o to Claude 3.5 Sonnet

- Category: Model swap
- Review state: Verified complete
- Source: `src/agents/support/model.ts`
- Evidence: Evaluation score and latency comparison attached

## PR #191: Added SQL Execution Tool to Support Agent

- Category: Agent tool
- Review state: High impact, requires review
- Source: `src/agents/support/tools.ts`
- Evidence: Permission scope expanded to `write:db_analytics`

Vouqis collects evidence. It does not certify compliance or AI safety.