# Change log (review pass)

The original submission is saved, unchanged, in `output/original_submission/`.

## Checks run (`output/review_check.py`)
- **CSV:** `schedule.csv` reads cleanly: 4 rows, and every row has the right number of fields.
- **UTC times:** every row's UTC time is its Pacific time plus 7 hours. In the first pass the script only converted the chosen start; the other UTC times were done by hand. All of them are now checked.
- **Feasibility:** I re-checked at 1-minute steps; the first pass used 5-minute steps. There are 12 feasible options, the earliest is 14:00 in Room C, and that matches the plan. The plan itself hasn't changed.
- **Risk note length:** under 500 words (the script prints the exact count).
- **Invitation:** it includes the required people, the optional attendee, the room, the agenda and the NOT SENT marker.

## Corrections
1. `schedule.csv`: Theo's prep location said "Theo's own workspace", which the inputs don't say. It now reads "Not specified in inputs (no room required)".
2. `risk-note.md`: the backup options were listed at 5-minute steps. They now give the full feasible ranges: Room C starting 14:01–14:05, Room B starting 14:10–14:15, and Room B rejected for starts 14:00–14:09. The note now names the 1-minute check script.
3. `risk-note.md`: the "one decision" section combined two questions. It's now one: who the security owner is. Whether they attend or send their decision in writing follows from that.
4. `invitation-draft.txt`: the agenda's timings and presenters are now marked as a drafter's suggestion for Maya to confirm. The inputs don't specify them.

## Not changed (checked, no error found)
- **The plan:** 14:00–14:45 PT in Room C, Theo's prep 13:30–14:00, setup 13:50–14:00, remote follow-up 14:45–15:00.
- **Morning rejection:** it still rests only on room availability and Maya's 11:00 cutoff. No constraint was relaxed.
- **Security owner:** still unresolved, and still not assumed to be Maya (`update.txt`).
