[ReneFr81]

What Changed

Add an authenticated goal-control endpoint for reading, pausing, and explicitly resuming an existing native Codex goal. Requests must match the persisted provider binding and live native session. Status updates preserve the objective, budget, and usage history; stale owners, active turns, and repeated mutation attempts fail with visible diagnostics.

Why

T3 currently exposes no route to native goal control, so sending another turn does not resume a usage-limited goal. Codex owns asynchronous continuation after resume. This change sends no additional wake and distinguishes status acceptance from observed continuation. Native quota denials remain visible and are never automatically reset.

This source-only PR contains one repair commit on a non-default integration branch rooted at the reviewed nightly baseline. Distribution and service activation remain separate decisions.

Validation

• 217 tests passed across eight focused files, covering authorization, routing, accounting preservation, active turns, duplicate attempts, and quota errors.
• A disposable built-server HTTP exercise verified get/pause/resume, incorrect-owner rejection, duplicate rejection, one native continuation, and a loopback quota denial. No subscription credentials were used.
• Scoped server typecheck and lint passed. Independent Astra-high review passed at this exact base and head, with no actionable findings.

Visual evidence: not applicable; server and contracts only, with no rendered UI changes.

Checklist

• One concern: native goal-control transport and regression coverage
• Explained behavior and cause
• Validated source and disposable assembled-server behavior

Implementation and independent review used GPT-6 Astra through 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	—	13.6 KiB	—	15.1 KiB	✅
Codex	Thread snapshot wire	—	7.0 KiB	—	7.3 KiB	✅
Codex	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Codex	Live turn WebSocket decoded	—	57.0 KiB	—	66.4 KiB	✅
Codex	Live turn messages	—	9	—	21	✅
Claude	Total thread wire	—	13.6 KiB	—	15.1 KiB	✅
Claude	Thread snapshot wire	—	7.1 KiB	—	7.3 KiB	✅
Claude	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Claude	Live turn WebSocket decoded	—	57.8 KiB	—	66.4 KiB	✅
Claude	Live turn messages	—	9	—	21	✅

Baseline: unavailable · PR result: 5b2d183 · 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: 113.9 KiB
• Claude decoded thread snapshot: 114.6 KiB

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