How I Built a File-Timestamp-Based Feedback Loop to Enforce AI Output Quality

The problem: AI outputs are probabilistic, and prompts have a ceiling LLMs produce probabilistic outputs. No matter how good your prompt is, edge cases will fail — hallucinations, omissions, format dr...