[unlockdigital]

Problem

Usage could show a full Copilot allowance for an exhausted account and leave a configured Claude API balance unknown. Undisclosed credit amounts and connections without usage reports also lacked useful explanations.

Change

• Preserve exhausted Copilot allowances when hasQuota is false, exclude unlimited chat/completion entitlements from premium-credit percentages, and omit expired reset timestamps.
• Make the saved-profile Console billing reader available when provider services are created, and retain explicit failure status after unsuccessful quota refreshes.
• Explain credit availability and missing usage on each account card. Keep billing links in account details and remove the duplicate billing sections.

Scope and approval

Maintainer-requested usage reliability update for the Unlock Digital fork. The shared provider fixes apply across clients; the account-card changes apply to web and desktop.

Verification

• vp run check:changed origin/unlock-digital/main: formatting, lint, both knip passes, GraphQL generation, typechecks for 33 affected packages, focused tests, and React compiler checks.
• Regression coverage includes exhausted Copilot credits, stale quota reads, credit amounts withheld by the provider, missing usage explanations, and billing links in account details.
• Live local verification confirmed the expected exhausted-credit state and a numeric API balance after repeated refreshes.
• Gitleaks found no secrets in the changed source files. Runtime settings, credentials, cookies, real account identifiers, and live balances remain local. These captures use synthetic accounts and balances; review assets are hosted outside the source commit.

Before:

After:

Model: GPT-6.1 Sol. Harness: Codex in T3 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	+6 B (+0.1%)	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	1.3 KiB	+6 B (+0.5%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.6 KiB	21.6 KiB	0 B (0.0%)	24.4 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	1.3 KiB	1.3 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 165192e · PR result: 73a4c51 · 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.