fix(web): keep an agent view visible instead of folding it away

[tusharbhardwaj-bk]

Reported after testing #141: agent views were not reliably visible in chat.

What was wrong

A t3_show_ui row arrives as an ordinary MCP tool row, so the work log treated
it like one. Two separate mechanisms hid it:

• Turn fold. Once a turn settles, its tool rows fold away entirely. This was
  the dominant one — a finished turn showed no view at all.
• Work-group toggle. In a longer run, tool rows collapse behind
  "+N tool calls".

Either way the box the agent drew was present in the data and invisible to the
reader.

The fix

Agent-spawn CTA rows already carry exactly this exemption, with the same
rationale — a still-running fleet must not hide behind a toggle. Rather than add
a second special case, the existing agentSpawn checks generalise into one
predicate, workEntryStaysVisible, used at all six sites that fold or collapse
work rows.

A view is the answer the agent produced, not a record that work happened. Folded
away it may as well not have been drawn.

The inline box already defaulted to open, and stays that way — this changes only
whether the row survives grouping. Expand-over-transcript from #141 is
unaffected.

Verification

New MessagesTimeline.agentUi.logic.test.ts — 6 tests:

• a view survives a run of tool calls without the reader expanding anything
• control: the same run without a view still folds away completely, so the
  collapse behaviour is intact for ordinary tool calls
• several views in one run all stay visible
• a lone view renders on its own
• the predicate covers spawn rows and views, and leaves ordinary rows collapsible

Existing MessagesTimeline.logic.test.ts passes unchanged (53 together). Wider
chat + fork web suites: 523 tests green. tsgo clean, format clean, fork markers
pass.

Written by Claude Opus 5 in T3 Code.

🤖 Generated with Claude Code

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[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	—	13.5 KiB	—	15.1 KiB	✅
Codex	Thread snapshot wire	—	7.2 KiB	—	7.3 KiB	✅
Codex	Live turn WebSocket wire	—	6.4 KiB	—	7.8 KiB	✅
Codex	Live turn WebSocket decoded	—	55.6 KiB	—	66.4 KiB	✅
Codex	Live turn messages	—	10	—	21	✅
Claude	Total thread wire	—	13.6 KiB	—	15.1 KiB	✅
Claude	Thread snapshot wire	—	7.2 KiB	—	7.3 KiB	✅
Claude	Live turn WebSocket wire	—	6.4 KiB	—	7.8 KiB	✅
Claude	Live turn WebSocket decoded	—	56.5 KiB	—	66.4 KiB	✅
Claude	Live turn messages	—	11	—	21	✅

Baseline: unavailable · PR result: 4fd50c7 · 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: 110.4 KiB
• Claude decoded thread snapshot: 111.1 KiB

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