Short comma-separated key hints everywhere; write down what is left to build

.pi/memory-vault/packs/paperclip-harness/60-observations/tmp-repository-structure-and-configuration-patterns.md

@@ -2488,3 +2488,69 @@ Choosing a task here writes the same choice the Board writes, so `Reply>reply` a
 - runbook: [[packs/paperclip-harness/70-runbooks/tmp-deployment-operations|navigate tasks and runs inside Work>work and merge safely (paperclip-harness)]]
 - session: [[packs/paperclip-harness/80-sessions/rich-discovery-home-2026-11-11-02-17-44|Rich discovery: home 2026-11-11 02:17:44]]
 - context: [[packs/paperclip-harness/20-context/home|home]]
+
+
+---
+
+## Update (2026-11-11 02:28:06)
+
+## Structural/configuration observations (2026-11-11 02:28:06)
+
+Each of those was a real defect, not misunderstanding on your part.
+
+## Runs were not selectable — they are now
+
+The cursor existed but nothing drew it, so the list looked inert. The chosen run is now marked and highlighted:
+
+```
+╭─ the work · 2 runs   runs   change   facts   review   log   merge ──·focus·╮
+│ state              woken by     started    ended      took
+│ ▸ succeeded        automation   15:56:28   15:56:59   31s
+│   succeeded        assignment   15:55:40   15:56:26   47s
+```
+
+`↑`/`↓` moves, `Enter` opens that run's log. The tab title now says how many runs there are, so `2 runs` versus `1 run` is visible without guessing.
+
+## The log no longer says "choose a run" when you cannot
+
+When a task opens, its newest run is chosen for you, so the log tab already has something in it. `Enter` on a different run replaces it.
+
+## The merge tab was showing another task's merge
+
+That was wrong and misleading. It now says plainly:
+
+```
+no merge has been run from this screen for AUD-166
+
+the last merge recorded was for another task:
+merge of AUD-171 started 12:53:39
+
+press m to merge this task
+```
+
+## Why a file-only task gets merged and rebased
+
+Because that file is a **commit**, and the commit sits on the task's branch, not on `main`. Merging is the only step that puts it on `main`. The rebase is not a rewrite of your work — it moves the branch onto the current tip of `main` so the checks run against what would actually land. If `main` has not moved since the branch was cut, the rebase does nothing and appears as a single line in the log.
+
+`AUD-171` created `STATUS-2.md` and merged it: that is exactly what the merge did — put that one file on `main`, at `a784e0f`. Nothing else was touched, and the review found nothing to look at because a new two-line file gives a reviewer very little.
+
+## What is still true about AUD-166
+
+It has two runs (`assignment`, then `automation` — Paperclip's corrective wake), no review recorded because it was never merged, and a diff. So: runs tab shows two rows, `Enter` opens either log, `change` shows the diff, `review` says none recorded, `merge` now says none has been run for it.
+
+## Navigation you can use without the Board
+
+In `Work>work`: `n` / `p` for the next or previous task (`task 6 of 13` in the header), `o` to pick from a list, `[` / `]` for tabs. Choosing a task there also updates `Reply>reply` and the rest, because it writes the same choice the Board writes.
+
+Committed as `f949345`; `docs/RUNS-AND-MERGE.md` now carries the explanation of runs, the merge order, and why a one-file task goes through the same gate.
+
+## Evidence
+
+- Confidence: 82%
+- Reason: Captured factual repo structure/configuration patterns from rich discovery
+
+## Related learned memories <!-- same-turn-2026-11-11 02:28:06 -->
+
+- session: [[packs/paperclip-harness/80-sessions/rich-discovery-each-of-those-was-a-real-defect-not-misunders|Rich discovery: each-of-those-was-a-real-defect-not-misunderstanding-on-your 2026-11-11 02:28:06]]
+- context: [[packs/paperclip-harness/20-context/each-of-those-was-a-real-defect-not-misunderstanding-on-your|each-of-those-was-a-real-defect-not-misunderstanding-on-your]]
+- observation: [[packs/paperclip-harness/60-observations/tmp-repository-structure-and-configuration-patterns|each-of-those-was-a-real-defect-not-misunderstanding-on-your repository structure and configuration patterns]]