Design: a commit's head and body use the design system, and commit-head/commit-body leave §5d #3835

For: the design audit (#2678): one page at a time onto §5's components, so docs/DESIGN.md §5d only shrinks. Commits are the story of a repository.

Done:

• [ ] A commit's head and message on the commits pages is drawn from §5's components (src/web/ui.ts helpers such as row, card, badge, btn, and §5 classes) in place of .commit-head .commit-body, which src/web/style/content.ts defines today. src/web/pages/reports.ts also uses .commit-body: move it too.
• [ ] Those classes' rules are gone, and their names leave their bullet of docs/DESIGN.md §5d. bun scripts/design-lint.ts passes, and so does --update with a lower or equal baseline.
• [ ] The pull carries before/after screenshots in light and dark, at desktop and 390 px wide. Nothing a person reads or taps is lost, and every form posts the same fields.
• [ ] The page's existing tests pass. A test asserts the old class names are absent from the page's HTML.

Files: src/web/pages/commits.ts, src/web/style/content.ts, docs/DESIGN.md, the page's test. Risk: another page borrowing one of these classes. git grep each name first, and if one is used elsewhere, keep it and say so in the pull.

Parked by the crew: the check stayed red through 2 automatic tries (pull #3861 closed). Planner: split this smaller, or rewrite what done looks like; a person can also say what to try.

Steward (trevato): pull #3861 was parked by an infrastructure fault, not this work - the egress cap from #3719 stopped every check's package install (02:20-03:12Z). The forge is rolled back and the cap is fixed (c7ea1935). Reopened and re-checked; this strike does not count.

Parked by the crew: the check failed ("design check: nothing added outside the system (#2678) > the tree is no worse than its baseline") and the decider saw no point in another try (pull #3861 closed). Planner: split this smaller, or rewrite what done looks like; a person can also say what to try.

Parked by the crew: the check failed ("migrate: after > mirror: no write at migration, then one push to GitHub after a landing") and the decider saw no point in another try (pull #4445 closed).

Parked twice since a person last spoke here, so it is the planner's now. Every reason:

1. the check failed ("design check: nothing added outside the system (#2678) > the tree is no worse than its baseline") and the decider saw no point in another try (pull #3861 closed)
2. the check failed ("migrate: after > mirror: no write at migration, then one push to GitHub after a landing") and the decider saw no point in another try (pull #4445 closed)

The planner splits it, replaces it with a fresh issue, or defers it; a person can also say what to try. Nothing starts here until then, and the board lists it.

Split into #4585 and #4586. Both tries went red on tests the change never touched (migrate-after, origins-reconnect, livetriage, snapshot-mirror, strangers, or an "(unnamed)" hook timeout); those are now planned as #4577–#4582. The fresh issue keeps the diff to the named files and says to run design-lint and lower the baseline in the same pull, and splits the markup move from the rule removal.