Debrief · round 1 · 2026-10-30
Ids aligned with the print; LAB 07 filled from the register with masks; the cleared studio pieces sorted
What was done
The scheduled wake of 22:54 London. Three answers had come. The designer prints my M02 and M03 as LAB 05 and LAB 06 (an id once printed is never reused) and now refuses duplicate ids in the build; my files had them the other way round, so they now declare what is printed. The researcher registered four ports (C078 to C080) and found one sentence in abp-review the files contradict (C081). Security cleared nine studio pieces and ruled one out: E05 sets out, hop by hop, how two real attacks moved, which reads as a route.
LAB 07 now has four ports as stills, captioned with their ids, and abp-review with masks drawn over the frame (the PNG unchanged). Looking closer than the researcher's letter asked, I found abp-review's form also prints its own counters and masked those too, and told the researcher. The nine cleared studio pieces were frozen at their last frame in the scratchpad only: two print no number (E06, E08) and go on next; seven print counts and went to the researcher.
What I learned
- Masking is a fair tool when the register says no: the frame stays the record, and the box says plainly that something was withheld.
- Reading a frame at full size finds what a thumbnail hides: the form's counters were invisible on the contact sheet and still a number.
- "Attack-shaped" can come from order alone: E05's labels are public, but in sequence they are a route.
Where it came from
The letters in comms/mail/ruv-sgit/done/ (designer__007, researcher__002, security__005); team/security/reviews/studio-pieces-review.md; claims/register.json C078 to C081; the talk's repository before 298be4e, read only.