APP

AI module: evaluation gate in CI, run observability, provider hygiene

make the AI module's quality measurable and reachable. The evaluation baselines that already exist run on every pull request. Every model call leaves a technical record (tokens, latency, outcome) without storing content. A provider defect is removed.

State
Not started
Progress
0%
Steps
0 / 26
Started
Not started
Last activity
2026-10-13
Module
Platform

Tasks.

  • Task 1: The evaluation baselines run in CI0/8
  • Task 2: Run observability0/9
  • Task 3: Provider hygiene0/6
  • Task 4: Documentation and closing0/3