feat(settings): port PR-aware inactivity settlement to orchestration v2

[tris203]

Reopens #12258 against t3code/codex-turn-mapping. Inactive exploratory threads should still settle automatically while users can keep PR-linked work active by choosing Threads without a PR. The other choices remain All threads and Off; defaults and independent merge/close rules are preserved.

The new base replaces the original settlement reactor with the V2 settlement service, so the earlier implementation cannot be carried over unchanged. This port applies the exclusion to V2's projection-backed decisions and includes scope in the settings key so environment and project changes trigger an immediate sweep. It retains the web/desktop and mobile controls, project overrides, shared settings sync, and capability filtering for older servers.

Validation: 316 focused tests passed, including V2 settlement/projection coverage, contracts, shared settings, and web/mobile scoped writes. Server, web, mobile, contracts, and client-runtime typechecks passed. Targeted lint passed with existing warnings; formatting and diff checks passed. An isolated dev server was provided for manual testing; no automated browser or simulator verification was performed.

UI reference from #12258 (the same controls are carried forward; this is not a new screenshot of the V2 build):

Model: GPT-6. Harness: Codex.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — This PR adds a defaulted inactivity-settlement scope and coordinates new production behavior across orchestration, settings contracts, project overrides, shared synchronization, and web/mobile controls. The cross-cutting runtime impact and product-default change require human review.

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

[juliusmarminge]

Note

This comment is posted by Julius' dot

Closing for incomplete UI evidence. At 86ac45f, the inactivity switch becomes an Off / All threads / Threads without a PR selector on web and mobile. The linked image from #12258 shows the resulting web dropdown only; neither PR supplies a before view or the changed native control. The focused tests and explicit V2-port limitations are useful, but do not complete the required before/after comparison.

Verification requires clear before/after screenshots for UI changes. Please add that comparison for the affected web and native controls, identify the tested build/environment, and request reconsideration. The PR-scoped inactivity option can be assessed as configuration of existing settlement behavior; this closure is for the missing UI evidence.