Show each annotation as its own chat message with element context

[shaunandrews]

Related issues

• Related to STU-2183
• Extracted from Design Exploration: Improve visual annotation workflows #4411. Independent of the other extractions.

How AI was used in this PR

Claude Code ported the transcript changes from #4411 onto trunk and adapted the element summary to trunk's annotation shape. Reviewed and tested manually in the Desktop app.

Proposed Changes

• Submitting preview annotations used to show a single "2 annotations submitted" bubble in the chat. Each note now appears as its own user message, with a chip underneath naming the element (h1.hero-title) and its nearby text.
• The agent prompt is unchanged; only the transcript display gains the per-note summaries, which are persisted on the user-prompt entry so they survive reload.
• The summary travels through every backend that starts a run (Desktop IPC, studio ui server, hosted) and is validated on the way in.

⚠️ Visual change: needs human review in light + dark mode.

Screenshots

Light	Dark

Testing Instructions

• Open a site preview, click Annotate, save two notes on different elements, click Submit.
• The chat shows two user messages, each with an element chip under it. Reload the session; they persist.
• Check the chip in light and dark.
• npm test -- apps/ui/src/ui-classic/components/session-view apps/ui/src/data/queries/use-agent-run.test.tsx apps/cli/commands/ai/sessions/tests/resume.test.ts packages/common/ai/visual-annotations.test.ts

Pre-merge Checklist

• Have you checked for TypeScript, React or other console errors?

🤖 Generated with Claude Code

[wpmobilebot]

📊 Performance Test Results

Comparing 6423c3f vs trunk

app-size

Metric	trunk	6423c3f	Diff	Change
App Size (Mac)	1458.03 MB	1458.04 MB	+0.01 MB	⚪ 0.0%

site-editor

Metric	trunk	6423c3f	Diff	Change
load	1171 ms	1209 ms	+38 ms	⚪ 0.0%

site-startup

Metric	trunk	6423c3f	Diff	Change
siteCreation	7513 ms	7526 ms	+13 ms	⚪ 0.0%
siteStartup	3402 ms	3389 ms	13 ms	⚪ 0.0%

---

Results are median values from multiple test runs.

Legend: 🟢 Improvement (faster) | 🔴 Regression (slower) | ⚪ No change (<50ms diff)