feat(web,mobile): show worktree setup progress instead of "Thinking"

[bk-agent-01]

Problem

When the first prompt of a new worktree thread is sent, the server creates the worktree and runs the setup script (often a multi-minute install) before the agent receives the prompt. During all of that, the chat showed "Preparing for 40s" plus a "Thinking" row, as if the agent were already working. The only hint of setup was a one-line "New worktree is being prepared" banner. The durable bootstrap already records worktreePhase / setupPhase on the execution intent; the UI just didn't use them.

Fix

• Shared (@t3tools/client-runtime/state/workspace-preparation): deriveWorkspacePreparation(intent) turns the intent's bootstrap phases into a worktree → setup → agent checklist, a one-line label, and an agentStarted flag. Returns null for turns that prepare nothing, so ordinary turns are untouched.
• Web: while preparing, the working row is an expandable "Setting up workspace · Running setup script…" row. Expanding it shows each step with a check, a Show output button for the setup terminal, and the failure detail if any. Once the agent takes over it collapses to "✓ Workspace ready" above the normal timer. "Thinking" is held back until the agent turn is actually running. The old one-line banner is hidden while the row is on screen and still shows failures.
• Mobile: the floating status pill names the current step ("Creating worktree…", "Running setup script…", "Starting agent…"), and the feed holds back the Thinking row until the agent starts.
• Server: prepareDurableBootstrap refreshes the execution snapshot when the setup step starts. Before this, step changes inside one preparation pass were never published, so clients could not see "worktree created, setup running" live.

The agent still waits for the setup script, as the durable bootstrap already did. No behaviour change there.

Verification

• packages/client-runtime/src/state/workspacePreparation.test.ts: 7 cases covering every phase combination.
• ProviderCommandReactor.test.ts: "publishes the running setup step before the setup script finishes". It fails without the server change (expected ['pending','pending'] to include 'running') and passes with it.
• Existing MessagesTimeline.test.tsx and thread-bootstrap tests pass. Typecheck passes for client-runtime, server, web and mobile. Fork-marker check passes.
• Screenshots: to follow from expbkt3 after deploy.

🤖 Generated with Claude Code

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ No successful main baseline artifact is available yet. This run establishes the initial measurement.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	—	13.8 KiB	—	15.1 KiB	✅
Codex	Thread snapshot wire	—	7.3 KiB	—	7.8 KiB	✅
Codex	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Codex	Live turn WebSocket decoded	—	57.1 KiB	—	66.4 KiB	✅
Codex	Live turn messages	—	9	—	21	✅
Claude	Total thread wire	—	13.8 KiB	—	15.1 KiB	✅
Claude	Thread snapshot wire	—	7.3 KiB	—	7.8 KiB	✅
Claude	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Claude	Live turn WebSocket decoded	—	57.9 KiB	—	66.4 KiB	✅
Claude	Live turn messages	—	9	—	21	✅

Baseline: unavailable · PR result: 5e25b20 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 114.9 KiB
• Claude decoded thread snapshot: 115.6 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.