Files

Sylpheed port agent a128642990 port: recover uncommitted GP_SAVE_LOAD.pak addition after restart

Session was OOM-killed mid-edit; this change was already correct and
self-documented in the why-block, just never committed.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-09-04 13:32:49 +00:00

audio.json

port: audit the remaining multi-leg claims -- the pattern I predicted is not there

2026-08-31 04:13:41 +00:00

flow.json

port: OPTIONS is reachable from the main menu

2026-09-03 19:16:42 +00:00

README.md

monorepo: one repository for the decoders, the port and the corpus

2026-08-29 11:34:46 +02:00

rendering.json

port: the rate's independent leg survived a challenge, and a failure mode recurred

2026-09-02 21:11:31 +00:00

screen_names.json

port: recover uncommitted GP_SAVE_LOAD.pak addition after restart

2026-09-04 13:32:49 +00:00

timing.json

port: withdraw "a third independent route", and measure the dwell residual I had only inferred

2026-09-02 15:10:40 +00:00

README.md

Authored decisions

Everything here is a decision we made, not something the disc said. It is hand-written, committed, and survives a re-export — unlike export/, which is regenerated wholesale and must never be hand-edited.

Every entry carries a why. When the RE agent decodes the real answer, delete the entry and let the exporter emit it; that deletion is the measure of progress.

See docs/FORMAT.md for the schemas and docs/BLOCKED.md for which HANDOFF question each placeholder is standing in for.