# E1: Formalizing: audit

Model: gpt-6-astra. Effort: a: high, b: high, c: medium, d: high, e: high.

Use one leg at a time in a study. Each leg continues the previous native checkpoint. Reader jobs and separate studies can run in parallel. Closure automatically validates, archives and dispatches. Repeat counts and pilot rounds below are starting examples.

```bash
python3 r-place/aggregator/handler/codex/run.py step STUDY E1a --start
python3 r-place/aggregator/handler/codex/run.py step STUDY E1b-01 --start
python3 r-place/aggregator/handler/codex/run.py step STUDY E1c --start
python3 r-place/aggregator/handler/codex/run.py step STUDY E1d --start
python3 r-place/aggregator/handler/codex/run.py step STUDY E1e --start
```

Talk through the live panel or `run.py chat ACTOR`. Open a checkpoint as an interview without changing the study. See [OpenAI operation](../../OPENAI.md).