fix(server): keep large tool results from resetting live streams

[rynfar]

A raw 9 MiB tool result could overflow an otherwise empty live-update buffer before its small client projection was applied. Both the thread and sidebar streams then failed with “The live event buffer is full,” disrupting updates regardless of the provider.

Project thread activity before retaining it, and keep only event identity in the shell queue before refetching its current aggregate. Preserve item/byte limits, sequence ordering, completion markers, replay and genuine overflow recovery. Persisted tool output and provider execution are unchanged. Selectively adopts the live-stream portion of pingdotgg/t3code#9799, source ce4712d5b04fb998f79fe132245289191147e5d5; the ledger identifies the portions outside this PR.

Validation: the two new real-WebSocket regression cases failed against the baseline. After the fix, all 183 tests across routing, live budgets, coalescing and payload projection passed. Scoped server typecheck, changed-file lint, formatting and diff checks passed. No visible layout or native changes require screenshots.

Model: GPT-6 Astra. Harness: Codex.

---

^{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.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.6 KiB	13.6 KiB	−7 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	6.6 KiB	6.6 KiB	−9 B (−0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	56.5 KiB	56.5 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	8	8	0 (0.0%)	21	✅
Claude	Total thread wire	13.8 KiB	13.8 KiB	−5 B (−0.0%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	+8 B (+0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.7 KiB	6.7 KiB	−13 B (−0.2%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	58.2 KiB	58.2 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: 71b11a4 · PR result: 88d92d7 · 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: 111.2 KiB
• Claude decoded thread snapshot: 111.9 KiB

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