Keep the personal half of the desktop in one place, and ask before installing it

Agent instructions, skills, SSH host aliases and expansion triggers are worth
having identical on every machine one person owns, and belong in none of the
shared configuration. They live in user/ now, with a manifest saying where each
piece goes and a link-user stage that puts it there.

That stage does nothing unless the machine said yes. Somebody who clones Panama
to try the desktop keeps their own ~/.claude/CLAUDE.md exactly where it was;
the question names the destinations and defaults to no. Anything displaced goes
to config/old rather than being deleted.

~/.claude/CLAUDE.md and ~/.codex/AGENTS.md were byte-identical copies of one
file, which is the drift this exists to prevent.

Also adds the vitals toggles for the battery and Claude usage readouts, which
had preferences and no way to reach them.

user/agents/skills/resolving-merge-conflicts/SKILL.md

@@ -0,0 +1,35 @@
+---
+name: resolving-merge-conflicts
+description: Use when a git merge or rebase is already in progress and conflicted, and hunks need resolving before it can finish.
+---
+
+# Resolving Merge Conflicts
+
+Resolve by **intent**, traced to each side's primary source, rather than by picking lines that look
+right. **Always finish the operation. Never `--abort`.**
+
+1. **See the current state.** Which operation is in flight (`git status`), which files conflict, and
+   what the two sides are (`git log --oneline HEAD..MERGE_HEAD`, or the rebase's `onto`).
+
+2. **Find the primary source for each side.** Understand *why* each change was made and what it was
+   for: read the commit messages, the merge request, and the ticket behind it. For KACP work the Jira
+   story is the primary source, and `.claude/docs/epics/` may already hold it. A hunk resolved without
+   knowing what either side was trying to do is a guess wearing a resolution's clothes.
+
+3. **Resolve each hunk.** Preserve both intents wherever they can coexist. Where they genuinely
+   conflict, take the one matching the stated goal of *this* merge and say so, noting the trade-off.
+   **Invent no new behaviour**: a conflict resolution is not the place for a third design neither
+   side asked for.
+
+4. **Run the project's checks.** Discover them rather than assuming: `package.json` scripts are the
+   source of truth. Typecheck, then tests, then format. Fix whatever the merge broke.
+
+5. **Finish the operation.** Stage everything and commit. On a rebase, continue until every commit is
+   replayed.
+
+## Done when
+
+- Every hunk was resolved against a primary source, not by appearance.
+- No behaviour exists that neither side had.
+- The project's typecheck and tests pass.
+- The merge or rebase is completed, never aborted.

user/agents/skills/resolving-merge-conflicts/agents/openai.yaml

@@ -0,0 +1,3 @@
+interface:
+  display_name: Resolving Merge Conflicts
+  short_description: Work an in-progress merge or rebase hunk by hunk, resolving by intent.