feat(ui): add SPA-side support for WebView2 native bridge

[AlexAlves87]

Summary

Wires the WebView2 bridge into the SPA lifecycle with a minimal, parity-aligned surface.

• app-native-bridge.ts — reduces NativeBridgeMessage to draft-text (inbound) and ready (outbound handshake). Removes recording-start/stop and voice-start/stop, since there is no corresponding SPA handler or UI surface today, and recording follows the parity decision in openclaw-windows-node#159. NativeBridgeHost now requires only handleChatDraftChange(next).
• app.ts — imports initNativeBridge, calls it in connectedCallback, and cleans it up in disconnectedCallback.
• draft-text goes through handleChatDraftChange so native-injected text resets input-history navigation the same way a real user edit does.

Context

The native side landed in openclaw-windows-node#192 (c7630fa) with origin validation, dispatcher marshaling, closed-window guards, sanitized logging, and payload JSON validation. This PR closes the SPA side of that minimal bridge surface.

Test plan

• In a WebView2 host: send {"type":"draft-text","payload":{"text":"hello"}} → chat input updates and input-history navigation resets
• Disconnect/reconnect component → no listener leak
• Outside WebView2 (regular browser) → initNativeBridge no-ops without errors

[clawsweeper]

Codex review: keeping this open for maintainer follow-up; there is still a little grit to resolve.

Keep this PR open. Current main still has no Control UI SPA-side WebView2 bridge or lifecycle wiring, while the draft PR adds a focused bridge module, app hookup, and tests for the draft-text and ready handshake surface. The linked Windows host-side bridge has merged in openclaw-windows-node#192, so this remains a live implementation candidate rather than obsolete cleanup.

Best possible solution:

Keep this PR open for normal maintainer review while it is draft. The best path is to review the focused Control UI bridge, confirm the merged Windows host bridge with a WebView2 smoke, keep the runtime validation and lifecycle cleanup tests, and land it if maintainers accept SPA-side WebView2 support.

What I checked:

• current_main_no_spa_bridge: A targeted current-main search in ui/src/ui and docs/web finds no initNativeBridge, app-native-bridge, chrome.webview, draft-text, NativeBridge, or WebView2 Control UI implementation. (fd2625a16252)
• app_lifecycle_lacks_bridge_wiring: OpenClawApp.connectedCallback currently installs the global keydown handler, calls handleConnected, and starts web push state; disconnectedCallback only removes the keydown handler and calls handleDisconnected. There is no native bridge init or cleanup on current main. (ui/src/ui/app.ts:585, fd2625a16252)
• draft_change_hook_exists: The existing draft-change helper sets chatMessage and resets input-history navigation, which matches the PR's intended route for native-injected draft text. (ui/src/ui/chat/input-history.ts:121, fd2625a16252)
• chat_render_uses_draft_hook: The chat render path passes the current draft and routes user draft edits through state.handleChatDraftChange, so a native draft-text message would integrate with the existing composer path. (ui/src/ui/app-render.ts:2316, fd2625a16252)
• docs_do_not_document_current_bridge: The Control UI docs describe a Vite + Lit SPA served by the Gateway and using the Gateway WebSocket; the WebChat docs describe Gateway WebSocket chat behavior. Neither documents WebView2 native bridge support on current main. Public docs: docs/web/control-ui.md. (docs/web/control-ui.md:10, fd2625a16252)
• pr_diff_is_focused_candidate: The PR head adds ui/src/ui/app-native-bridge.ts, ui/src/ui/app-native-bridge.test.ts, and five lines of lifecycle wiring in ui/src/ui/app.ts. The bridge surface is inbound draft-text plus outbound ready, with runtime guards for malformed messages and cleanup tests. (ui/src/ui/app-native-bridge.ts:1, 2aefa119696e)

Likely related people:

• BunsDev: Remote history shows BunsDev authored the original Control UI dashboard shell and the later chat infrastructure slice that introduced chat/input-history.ts, directly adjacent to the draft handling and app/render surfaces this PR wires into. (role: original Control UI and chat-infrastructure owner; confidence: high; commits: 3bbbe33a1b91, c5ea6134d041, e3ad82d86df0; files: ui/src/ui/app.ts, ui/src/ui/app-render.ts, ui/src/ui/chat/input-history.ts)
• Ivocin: Commit 8200d87 specifically hardened WebChat input history behavior, including draft/navigation consistency, which is the exact SPA behavior the native draft-text bridge intends to reuse. (role: recent input-history maintainer; confidence: medium; commits: 8200d878a340; files: ui/src/ui/chat/input-history.ts, ui/src/ui/app.ts)
• steipete: Recent Control UI history includes browser-safety and browser realtime Talk work by steipete, both adjacent to adding a browser/native bridge and reviewing its trust assumptions. (role: recent Control UI browser-surface maintainer; confidence: medium; commits: 9e149519fed8, 04066d246abc, c6ebd99a461d; files: ui/src/ui/app.ts, ui/src/ui/app-render.ts, docs/web/control-ui.md)

Remaining risk / open question:

• The PR introduces a new native-message trust boundary. The SPA-side diff validates payload shape, and the linked Windows host-side bridge validates origin, but maintainers should still run a real WebView2 smoke before merge.
• This was a read-only cleanup review, so no local tests or WebView2 runtime smoke were run.

Codex review notes: model gpt-5.5, reasoning high; reviewed against fd2625a16252.