Phase 0 — Inventory

Audit AI-heavy journeys, decide what becomes a Restormel Testing goal versus unit or traditional E2E, and align on how Keys and secrets will participate in test runs.

You should finish this phase with:

  • A short list of user outcomes that depend on AI (prompts, tools, retrieval, model routing).
  • A clear line between “deterministic app logic” tests (stay in your existing runners) and “outcome goals” for Testing.
  • A plan for Keys: which logical model slots tests will need (primary, optional vision/judge).
  • No commitment yet to huge coverage—MVP targets 5–8 meaningful goals.

Existing stack · Next: Phase 1 — Install and configure

Agent prompts

Optional: use these with a coding agent to implement this phase in your repo in a safe, gated sequence. Expand only when you need them.

  1. List three AI-affected user journeys in our product; rank by regression risk.
  2. For each, note what “pass” means in user terms vs technical signals we can assert.
  3. Mark which journeys stay in Playwright/unit tests only and which merit a Testing goal.