refactor(provider-core): add provider-core and provider-testing packages

[juliusmarminge]

Providers could not live outside apps/server: the driver contract, the helpers drivers share, and every way a driver reached the server were server-internal. This adds the packages a provider package builds on. No behavior change.

@t3tools/provider-core

• ./client: ProviderClientDefinition (label, settings schema, environment fields, badge, plain-data icon) and a registry. Web settings, onboarding, and usage read it instead of a web-only table.
• ./server/*: the driver and instance records, the V2 adapter interface, selection transitions, snapshot, auth, text-generation, and maintenance interfaces, plus the helpers drivers share that have no server dependencies (ids, failure items, MCP session and tool presentation, process groups, status probes, update ownership, usage limits, text-generation prompts).
• ProviderHost: the one service a provider uses to reach the server (paths, settings reads and changes, background-work gate, attachment paths), implemented once by the server. Every driver now takes it instead of the server's BackgroundPolicy.

@t3tools/provider-testing: replay transcript decoding, the replay gate, the git checkpoint workspace, and a scoped ProviderHost for driver tests.

Server modules keep their services and re-export moved types. Effect service modules are PascalCase (ProviderAdapter, ProviderHost, IdAllocator); utility modules are camelCase. Service keys follow the new module paths (in-memory tags; nothing persists them).

---

Stack: #17299 → #17300 → #17302. Replaces the earlier 7-PR stack (#17282–#17288).

Done by Claude Opus 5.5 in Claude Code.

🤖 Generated with Claude Code

---

[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	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.9 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 5809487 · PR result: 52fbad9 · 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: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

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

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This PR performs a broad provider-platform refactor that rewires production provider construction, runtime services, and client metadata across hundreds of files. It also changes authentication-related interfaces, so the scope and sensitivity warrant human validation.

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

[coderabbitai]

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

• Commit to this branch
• Create a new PR

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

_{Comment @coderabbitai help to get the list of available commands.}