feat(auth): share provider sign-in flows across clients

[juliusmarminge]

Provider setup has separate account flows for each driver, which makes new authentication methods hard to support consistently across environments and clients.

Add shared authentication contracts and interactions, a provider credential store, and account controls for web, desktop, and mobile. Coordinate credential changes with provider session admission and cleanup. Provider adapters retain ownership of token exchange and native credential formats. Account actions reflect verified sign-in state, dedicated API keys live under Setup using the shared settings row, and empty Runtime sections are hidden. The provider sidebar keeps Codex and Claude visible and only shows other default providers after configuration; configured instances remain visible when disabled. Account emails use the existing masking control.

Validation: focused authentication, credential storage, session admission, RPC authorization, Antigravity, and settings tests passed. Scoped contracts, client runtime, server, web, and mobile typechecks passed. The integrated stack passed 231 focused tests.

Live sign-in and before/after UI evidence are pending. This PR remains a draft.

Implemented with GPT-6 in the Codex harness.

[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: 8bd3aa5 · Source CI: failure

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.