This repository has been archived on 2026-10-19. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
DB-cleanup/merge-output/flags/CHECKER-PROMPT.md
T
StellarCrow 41766c68a2 chore: final state before the move to gart/gart-databases
Commits the work done since 2026-10-01 that was still local: the four-step
label and flag passes, the triage and recheck files, the v1 exports of
2026-10-06 and 2026-10-07, the split dossiers, and the Lopp PR 234 batch
(scripts, agent prompts, reader and checker files, the flag re-read of
2026-10-17, the review workbook and the import set of the 57 cases added
to stats2 as 506 to 562). Raw HTML and node_modules stay out (.gitignore).

The stats2 material in use was copied to gart/gart-databases (stats2/) on
2026-10-17; README-START-HERE.md now says so. This repo is closed.
2026-10-17 10:47:37 +02:00

2.2 KiB

Flag check, second reader: confirm or reject each change

A first reader tested the violence flags and death count of every case against FLAG-RULES.md and changed some values. You judge each change. Only confirmed changes reach the database.

Working directory: ~/Documents/Claude/GART/03-research-and-publishing/DB-cleanup/merge-output

Files

  • FLAG-RULES.md: the tests. Read all of it first.
  • flags/in-NN.json: the cases, with current values, locked_by_ruling, obtained, outcome, summary and dossier path.
  • flags/out-NN.json: the first reader's result. Do not change it.
  • Write your verdicts to flags/check-NN.json. Write nothing else. Read files with the Read tool; never open saved tool-output copies under /home/disi/.claude or /tmp. Fetch nothing from the web. Run no git commands.

What to do

  1. For every case and every unlocked field where the first reader's value differs from current: read the summary, and the dossier where the quote comes from there or the change is from yes to no, and decide whether the new value passes its test in FLAG-RULES.md and the current one fails. Confirm or reject.
  2. Then go through every case once more and look for an unlocked value the first reader left wrong: one that plainly fails its test. Report those as missed. Ignore locked values.

Reject when the text supports the current value, when the change rests on silence in the summary while the dossier reports the thing, or when it is a matter of taste. When unsure, reject.

Output

A JSON array: one object per changed value, plus one per missed change.

{"id": 123, "field": "weapons", "verdict": "confirm"}
{"id": 77, "field": "theft", "verdict": "reject", "reason": "the dossier says the ransom was paid"}
{"id": 301, "field": "detention", "verdict": "missed", "proposed": "yes", "quote": "tied up the two clerks", "reason": "staff restrained"}

A reject needs a reason. A missed change needs the proposed value, an exact passage from the text and a reason; it is listed for a person and not applied. An empty array [] is a valid result. Valid JSON, straight quotes.

When you finish

Reply with one line: the batch number, changes confirmed, rejected, missed. No victim names in the reply.