refactor: extract duplicated exec, traversal and item-building logic into shared utilities

[devin-ai-integration]

What does this PR do?

Removes ~580 lines of duplicated logic by extracting shared utilities and rewiring every caller. No behavior or output changes.

New utilities:

• utils/exec.ts — execCommand(command, args, timeout) (spawn-based, no shell) + findExecutable(candidates); replaces 5 near-identical copies (docker, homebrew, dns-cache, purgeable, timemachine) and 3 copies of the absolute-path binary probe. Absolute candidate paths and arg arrays are preserved, so the $PATH/injection hardening is unchanged; per-call timeouts keep their original values (10s DNS, 60s purge, 30s default).
• utils/concurrency.ts — mapPool(items, concurrency, fn); replaces the private pool in git-worktrees and the Promise.race-based runWithConcurrency in scanners/index.ts, which now just wraps it to keep swallowing+logging per-scanner failures.
• utils/walk.ts — walkFiles(root, { maxDepth, includeHidden }, onFile); replaces the recursive depth-limited traversals in large-files and duplicates (hidden-entry skipping and permission-error swallowing preserved).
• utils/fs.ts — createPathItem(path, name, { skipEmpty }) and collectDirectoryItems(paths); replace the exists -> getSize -> stat -> push block repeated in 4 scanners and the if (await exists(dir)) items.push(...await getDirectoryItems(dir)) block repeated in 7.
• utils/size.ts — sumItemSizes(items) for the reduce((sum, i) => sum + i.size, 0) repetition.
• commands/shared.ts — SAFETY_ICONS (previously defined identically in clean, scan, interactive), plus runCleanSelections/countSelectedItems/sumSelectedSize for the clean loop duplicated between clean and interactive.

MaintenanceResult moved to maintenance/types.ts (it was declared twice and imported from dns-cache by timemachine); maintenance/index.ts re-exports it from the new location.

Print/formatting code in clean vs interactive was intentionally left alone — the wording differs, so merging it would change user-visible output.

Type of change

• Improvement to existing feature

Related issue

N/A

Checklist

• bun run lint passes
• bun run test passes (no regressions)
• bun run build succeeds
• New tests added for new behavior (utils/walk.test.ts, utils/concurrency.test.ts; 354 tests pass)
• I've tested this on macOS locally — not possible from a Linux dev box; covered by the existing suite

Notes for reviewer

Worth a look at runWithConcurrency in scanners/index.ts: the old version raced promises as it queued them, the new one is a fixed worker pool over mapPool. Same limit and same error handling, but scheduling order differs slightly.

Link to Devin session: https://app.devin.ai/sessions/f560eeb632a4468a93f3ce06d0b53cc5
Requested by: @guhcostan

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring