AI evals, authoring and running benchmark cases

ai_evals/ is a black-box benchmark runner for the Windmill AI generation modes: flow, app, script, cli, global. It always tests the current production prompts, tools, and guidance in this checkout. Each attempt runs the real production path, deterministic validation, then LLM judging.

The goal is to test current production guidance with realistic user requests, not to pin one exact implementation shape.

Running benchmarks

cd ai_evals
bun install                       # first time; frontend modes also need `cd frontend && bun install`
bun run cli -- models             # list model aliases
bun run cli -- cases global       # list cases for a mode
bun run cli -- run global global-test1-script-create --model sonnet

Frontend modes (flow/script/app/global) route model calls through a Windmill backend's /api/w/<ws>/ai/proxy, so you need any reachable backend:

WMILL_AI_EVAL_BACKEND_URL=http://127.0.0.1:<port> WMILL_AI_EVAL_BACKEND_WORKSPACE=integration-tests \
  bun run cli -- run global <caseIds...> --models sonnet,gpt-5.5,gemini-3.1-pro-preview

• Reuse an existing workspace. CE builds cap workspaces, so temp-workspace creation 400s ("reached workspace limit"). Always set WMILL_AI_EVAL_BACKEND_WORKSPACE=integration-tests (or any existing workspace) to reuse one. The only side effect of a run is upserting an f/evals/ai/<provider> resource there.
• Provider keys live in ai_evals/.env and are auto-loaded by bun. The judge is a separate Anthropic call (default claude-sonnet-4-6) regardless of the model under test.

Authoring core rules

1. Write prompts like a real user request.
2. Prefer behavior, inputs, constraints, and outcomes over internal implementation.
3. Keep deterministic validation narrow and hard.
4. Put semantic expectations in judgeChecklist.
5. Use expected fixtures only when exact structure really matters.