fix(web): keep the floating pill nav within the viewport

[maarco]

What Changed

• The docked pill's position along its edge is clamped from its measured size, so both ends stay 20px inside the window on all four docks. The clamped value is saved.
• The pill's size cap now comes from the viewport and the pill scale instead of its saved position, so the border wraps the whole pill.
• With the pill unlocked, dragging from the grip moves the pill. A drag no longer toggles the lock; a click still does.
• Workspace headers move down only while the pill overlaps the top header band, using the pill's real rectangle. A pill docked left or right never moves the header.

Why

A pill saved near either end of its edge (10% or 90%) was centred on that percentage whatever its width, so in a narrower window part of it sat outside the viewport. Its size cap also shrank toward those ends, so most controls were cut off. The header spacing only handled a top dock, so a pill dragged across the header could cover the title.

UI Changes

No screenshots are attached. Measured in a running app at 820px wide with the pill saved at 10% along its edge:

Dock	Before (main)	After
Top	pill 124px wide, most controls cut off	pill 709px wide, all controls visible, 20px from the left edge
Left	pill 140px tall, most controls cut off	pill 673px tall, all controls visible, 20px from the top edge

The bottom and right docks use the same clamp, and all eight dock and offset combinations (top, bottom, left, right at 10% and 90%) stayed inside the window with a 20px margin. At 390px wide the pill is the full-width top rail, before and after.

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for any UI changes (not attached; measurements above)
• I included a video for animation/interaction changes (n/a: no animation change)

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.5 KiB	13.5 KiB	−19 B (−0.1%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	+2 B (+0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.4 KiB	6.4 KiB	−21 B (−0.3%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	56.2 KiB	56.2 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	9	9	0 (0.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	+18 B (+0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	−6 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.4 KiB	6.4 KiB	+24 B (+0.4%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.0 KiB	57.0 KiB	+44 B (+0.1%)	66.4 KiB	✅
Claude	Live turn messages	8	9	+1 (+12.5%)	21	✅

Baseline: db6fa48 · PR result: 11c571e · 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: 113.9 KiB
• Claude decoded thread snapshot: 114.6 KiB

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