fix(server): preserve context occupancy across model handoffs

[Yash-Singh1]

Summary

• Switching models on the same native thread no longer drops the measured context size and replaces it with a byte-length estimate.
• The handoff keeps that occupancy, drops the previous model's compaction threshold, and uses the new model's window when it is known. A conversation that still fits is no longer refused with "Insufficient context allowance."
• A genuinely full window still blocks the handoff. A new native thread still starts with an empty occupancy.

Test plan

• vp test run src/orchestration-v2/ContextHandoffBudget.test.ts
• vp test run src/orchestration-v2/testkit/ProviderSwitch.integration.test.ts
• Switch models on a thread that previously hit a usage limit and confirm the next turn starts instead of reporting insufficient context allowance
• Branching the last turn of that thread still runs
• Switching onto a model whose window is actually smaller than the current occupancy still refuses the handoff

Made with Cursor

[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	—	4.9 KiB	—	6.8 KiB	✅
Codex	Thread snapshot wire	—	3.7 KiB	—	4.9 KiB	✅
Codex	Live turn WebSocket wire	—	1.2 KiB	—	2.0 KiB	✅
Codex	Live turn WebSocket decoded	—	20.4 KiB	—	29.3 KiB	✅
Codex	Live turn messages	—	2	—	8	✅
Claude	Total thread wire	—	4.9 KiB	—	6.8 KiB	✅
Claude	Thread snapshot wire	—	3.7 KiB	—	4.9 KiB	✅
Claude	Live turn WebSocket wire	—	1.2 KiB	—	2.0 KiB	✅
Claude	Live turn WebSocket decoded	—	20.8 KiB	—	29.3 KiB	✅
Claude	Live turn messages	—	2	—	8	✅

Baseline: unavailable · PR result: b892868 · 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: 106.1 KiB
• Claude decoded thread snapshot: 106.4 KiB

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

[macroscopeapp]

Approvability

Verdict: Approved at b892868

Macroscope's review found this PR approvable — This is a focused server bug fix that preserves measured context occupancy across existing model handoffs while resetting only stale capacity metadata, with new-thread behavior unchanged. The production change is localized and supported by unit and integration coverage for model switches, retries, telemetry, and native replacement.

^{You can add or adjust custom eligibility rules. Learn more.}