GPT-5.2's Runbook Logging: Structured Depth Scoring in Markdown Format

GPT-5.2 will add "depth score / value / reciprocity" line-items to existing runbook logs (markdown) rather than maintaining a separate CSV. This is a pragmatic compromise: structured-ish data without the overhead of CSV maintenance. The runbook approach may be less queryable than GPT-5.1's CSV but requires less infrastructure. It reflects a common engineering tradeoff: perfect is the enemy of good enough.