Publish live Grok tokens and xAI spend into Usage & Spend

[Chipagosfinest]

Summary

• Put enabled Grok and xAI into the shared Usage & Spend catalog that Overview already counts.
• Map xAI Management API daily spend (the existing billing chart) as vendor-metered USD. Prepaid ledger balance stays remaining credit, not spend.
• Map Grok ~/.grok/sessions/**/signals.json into per-day token buckets. SuperGrok/X Premium+ credits stay a quota window and are never converted into dollars.
• UTC day labels for xAI follow the same dashboard calendar mapping as Mistral and OpenRouter.

This continues the Overview spend work in #3023, #3054, #3063, and #3067: every enabled subscription should contribute the strongest truthful signal instead of silently dropping out.

Why this slice

#2759 / #2855 were closed so review could happen concern-by-concern. After #3067, Grok and xAI were still supportsTokenCost: false, so they inflated Overview coverage (x of y subscriptions have spend) without publishing live data.

Test plan

• GrokLocalSessionScannerTests — daily local buckets, no invented dollars, empty home publishes nothing, and local scan time owns Today
• XAICostUsageMappingTests — chart → vendor-metered catalog; prepaid balance alone is not spend; partial history stays estimated
• GrokXAISpendCatalogTests — both providers are snapshot-backed spend sources plus opt-in redacted live catalog proof
• XAIProviderTests — existing golden balance/chart behavior unchanged; malformed successful history stays unavailable
• GrokMenuCardModelTests — credit-bar presentation unchanged
• PopupLocalizationTests — OpenRouter generic details localize titles/row labels via L() after 84a4ca7, while values and chart points stay canonical

Exact-head verification

Verified on a4f80675a4d1a2d1d76beafd335be6e2a5267002 after rebase onto main and Peter's merge of 84a4ca725:

make check
# passed: SwiftFormat, SwiftLint, JS/TS lint, generated-resource and repository gates

swift test --filter PopupLocalizationTests
# 6 tests in 1 suite passed

swift test --filter 'PopupLocalizationTests|GrokLocalSessionScannerTests|XAICostUsageMappingTests|XAIProviderTests|GrokXAISpendCatalogTests'
# 29 tests in 5 suites passed on the prior head; PopupLocalizationTests re-verified on this head

Redacted after-fix output from the real local Grok scanner through SpendDashboardModel.build:

catalog_source=grok
today_tokens=4841945
last_30_days_tokens=4841945
daily_buckets=1
available_sources=grok

No paths, account identifiers, session identifiers, prompts, or model names are included in this proof.

Residual risk

• xAI still needs a Management API key + team ID for dollars to appear. Without them the provider stays enabled and reports unavailable, not $0.
• Grok token totals are local session signals, not the SuperGrok bill. That is intentional.

[clawsweeper]

🦞👀
ClawSweeper picked this up.

Pull request received. I will update this pull request when review starts.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 26fd232eef

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[clawsweeper]

Codex review: needs changes before merge. Reviewed August 20, 2026, 4:07 AM ET / 08:07 UTC.

ClawSweeper review

What this changes

The PR adds Grok local-session token activity and xAI Management API daily spend as sources for CodexBar’s shared Usage & Spend catalog.

Merge readiness

⚠️ Needs maintainer review before merge - 4 items remain

The prior cache-path blocker remains on the current head: Grok session files are still rescanned after refresh while constructing the shared catalog. The PR otherwise has adequate redacted real-behavior proof and no security concern.

Priority: P2
Reviewed head: a4f80675a4d1a2d1d76beafd335be6e2a5267002

Review scores

Measure	Result	What it means
Overall readiness	🦐 gold shrimp (3/6)	The feature has strong redacted local proof, but the unresolved cache-only P2 prevents merge readiness.
Proof confidence	🦞 diamond lobster (5/6)	Sufficient (terminal): Redacted terminal output from a real local Grok scanner through the shared catalog shows after-fix token and daily-bucket publication; it contains no account or session identifiers.
Patch quality	🦐 gold shrimp (3/6)	2 actionable review findings remain.

Verification

Check	Result	Evidence
Real behavior	Verified	Sufficient (terminal): Redacted terminal output from a real local Grok scanner through the shared catalog shows after-fix token and daily-bucket publication; it contains no account or session identifiers.
Evidence reviewed	6 items	Current Grok mapper rescans local files: The Grok branch calls the recursive local-session scanner instead of reusing the snapshot already built during the provider refresh. Catalog request invokes the mapper after refresh: The request builder refreshes snapshot-backed providers, then its Grok path calls the same mapper to populate catalog input; this retains the prior scan-on-catalog-path defect. Refresh already has a Grok-derived snapshot: The Grok provider converts its local summary to cost usage during the refresh flow, so a cached publication can supply the catalog without another scan.
Findings	2 actionable findings	[P2] Read Grok catalog data from the published snapshot [P3] Remove the release-owned changelog entry
Security	None	None.

Live Verification

Command: swift test --filter XAICostUsageMappingTests

Result: FAIL (failed) — execution before step 1 run: sh -lc pnpm install --ignore-scripts --frozen-lockfile failed: ! Corepack is about to download https://registry.npmjs.org/pnpm/-/pnpm-11.22.0.tgz

sh -lc pnpm install --ignore-scripts --frozen-lockfile failed: ! Corepack is about to download https://registry.npmjs.org/pnpm/-/pnpm-11.22.0.tgz

Assertions:

• FAIL expect_output: Test run

How this fits together

Provider refreshes produce token/spend snapshots that the shared catalog publishes to both the Overview menu and the Usage & Spend dashboard. Grok data comes from local session logs, while xAI data comes from the Management API’s daily billing history.

flowchart LR
A[Grok local session logs] --> B[Provider refresh]
C[xAI daily billing history] --> B
B --> D[Published provider snapshots]
D --> E[Shared Usage and Spend catalog]
E --> F[Overview menu]
E --> G[Settings dashboard]

Loading

Before merge

• Read Grok catalog data from the published snapshot (P2) - The prior P2 remains on this unchanged path: makeRequest refreshes Grok and then invokes this mapper again while capturing catalog input, which recursively walks ~/.grok/sessions. The refresh snapshot already carries costUsage; reuse its cached publication so menu/dashboard updates cannot rescan local logs on the main-actor catalog path.
• Remove the release-owned changelog entry (P3) - CHANGELOG.md is release-owned in this repository workflow, so this normal provider feature PR should leave release notes to the release process.
• Resolve merge risk (P1) - Opening or refreshing the catalog can synchronously walk a large ~/.grok/sessions tree on the main-actor request path, delaying menu/dashboard responsiveness and producing duplicate local file reads.
• Complete next step (P2) - A concrete cache-only repair and release-file removal can be made mechanically on this PR branch.

Findings

• [P2] Read Grok catalog data from the published snapshot — Sources/CodexBar/UsageStore+TokenCost.swift:477-479
• [P3] Remove the release-owned changelog entry — CHANGELOG.md:5

Agent review details

Security

None.

Review metrics

Metric	Value	Why it matters
Patch surface	19 files affected; 544 additions, 33 deletions	The change spans provider parsing, snapshot publication, dashboard aggregation, tests, and documentation.

Merge-risk options

Maintainer options:

1. Cache Grok session publication before merge (recommended)
   Reuse the Grok cost-usage value captured during refresh and add a regression proving catalog capture does not enumerate local session files.
2. Pause the provider extension
   Do not merge this slice if the cache-only publication boundary cannot be preserved without a larger lifecycle redesign.

Copy recommended automerge instruction

@clawsweeper automerge

Special instructions:
Reuse the Grok cost snapshot published during provider refresh for the shared catalog; do not call the local-session scanner from catalog capture or menu-facing mapping, and add focused regression coverage.

Technical review

Best possible solution:

Publish Grok’s scan once during provider refresh, retain it as the current provider snapshot/publication, and have catalog construction read only that cached value.

Do we have a high-confidence way to reproduce the issue?

Yes, from source: enable Grok with a large local session tree, trigger shared catalog capture, and the current mapper synchronously calls the local scanner after refresh.

Is this the best way to solve the issue?

No: the existing refresh snapshot already contains the Grok-derived cost usage, so rescanning during catalog capture is avoidable and violates the cache-only boundary.

Full review comments:

• [P2] Read Grok catalog data from the published snapshot — Sources/CodexBar/UsageStore+TokenCost.swift:477-479
  The prior P2 remains on this unchanged path: makeRequest refreshes Grok and then invokes this mapper again while capturing catalog input, which recursively walks ~/.grok/sessions. The refresh snapshot already carries costUsage; reuse its cached publication so menu/dashboard updates cannot rescan local logs on the main-actor catalog path.
  Confidence: 0.98
• [P3] Remove the release-owned changelog entry — CHANGELOG.md:5
CHANGELOG.md is release-owned in this repository workflow, so this normal provider feature PR should leave release notes to the release process.
  Confidence: 0.99

Overall correctness: patch is incorrect
Overall confidence: 0.98

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 84a4ca725d5d.

Labels

Label justifications:

• P2: The feature is useful but a main-actor local-file scan can cause a normal-priority responsiveness regression.
• merge-risk: 🚨 availability: Merging retains a recursive local-session scan on the catalog request path used by the shared menu/dashboard publisher.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦞 diamond lobster and patch quality is 🦐 gold shrimp.
• status: ⏳ waiting on author: ClawSweeper has contributor-facing work open and is waiting for author action. Sufficient (terminal): Redacted terminal output from a real local Grok scanner through the shared catalog shows after-fix token and daily-bucket publication; it contains no account or session identifiers.
• proof: sufficient: Contributor real behavior proof is sufficient. Redacted terminal output from a real local Grok scanner through the shared catalog shows after-fix token and daily-bucket publication; it contains no account or session identifiers.

Evidence

Acceptance criteria:

• [P1] swift test --filter GrokLocalSessionScannerTests.
• [P1] swift test --filter GrokXAISpendCatalogTests.
• [P1] swift test --filter XAICostUsageMappingTests.
• [P1] make check.

What I checked:

• Current Grok mapper rescans local files: The Grok branch calls the recursive local-session scanner instead of reusing the snapshot already built during the provider refresh. (Sources/CodexBar/UsageStore+TokenCost.swift:478, a4f80675a4d1)
• Catalog request invokes the mapper after refresh: The request builder refreshes snapshot-backed providers, then its Grok path calls the same mapper to populate catalog input; this retains the prior scan-on-catalog-path defect. (Sources/CodexBar/SpendDashboardController.swift:276, a4f80675a4d1)
• Refresh already has a Grok-derived snapshot: The Grok provider converts its local summary to cost usage during the refresh flow, so a cached publication can supply the catalog without another scan. (Sources/CodexBarCore/Providers/Grok/GrokStatusProbe.swift:69, a4f80675a4d1)
• Prior-review continuity: The immediately preceding completed review raised the same cache-only requirement, and blame ties the unchanged mapper to the current PR head. (Sources/CodexBar/UsageStore+TokenCost.swift:478, a4f80675a4d1)
• Shared-catalog provenance: The merged shared catalog predecessor is commit 4029705, which identifies the feature lineage this PR extends. (Sources/CodexBar/SpendDashboardController.swift:211, 40297051639b)
• Real behavior proof: The PR body provides redacted after-fix output from a real local Grok scanner through SpendDashboardModel.build, showing a Grok source, token totals, and one daily bucket. (a4f80675a4d1)

Likely related people:

• Chipagosfinest: The supplied history identifies this contributor as the author of the merged shared spend-source catalog work and of this provider-specific extension. (role: shared-catalog feature contributor; confidence: high; commits: 40297051639b, a4f80675a4d1; files: Sources/CodexBar/SpendDashboardController.swift, Sources/CodexBar/UsageStore+TokenCost.swift)
• Alec Gutman: Blame assigns the unchanged Grok catalog and mapper lines to the current PR head commit. (role: current-head source author; confidence: high; commits: a4f80675a4d1; files: Sources/CodexBar/SpendDashboardController.swift, Sources/CodexBar/UsageStore+TokenCost.swift)

Rank-up moves

Optional improvements that raise the rating; they are not merge blockers.

• Reuse the refresh-published Grok snapshot in catalog capture.
• Add focused coverage that catalog construction cannot trigger a Grok local-session scan.
• Remove the release-owned CHANGELOG.md edit.

Rating scale

Score	Internal tier	Crab rank	Meaning
6/6	S	🦀 challenger crab	Exceptional readiness
5/6	A	🦞 diamond lobster	Very strong readiness
4/6	B	🐚 platinum hermit	Good normal PR; ordinary maintainer review
3/6	C	🦐 gold shrimp	Useful, but confidence is limited
2/6	D	🦪 silver shellfish	Proof or implementation needs work
1/6	F	🧂 unranked krab	Not merge-ready
N/A	NA	🌊 off-meta tidepool	Rating does not apply

Overall follows the weaker of proof and patch quality.
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

Workflow

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

History

Review history (5 earlier review cycles)

• reviewed 2026-08-20T01:03:50.059Z sha 55e602c :: needs real behavior proof before merge. :: [P1] Preserve xAI analytics failures as unavailable | [P2] Publish local Grok tokens without a remote snapshot | [P2] Select xAI’s current UTC billing-day entry | [P2] Select Grok’s current local-day token bucket | [P2] Keep xAI’s 30-day history from posing as 365 days
• reviewed 2026-08-20T01:15:17.742Z sha bef86db :: needs real behavior proof before merge. :: [P1] Preserve xAI analytics failures as unavailable | [P2] Publish Grok tokens without a remote snapshot | [P2] Use xAI’s current UTC day for Today | [P2] Use Grok’s current local-day bucket for Today | [P2] Limit xAI coverage to its 30-day response
• reviewed 2026-08-20T01:36:58.552Z sha 4a93fbe :: needs real behavior proof before merge. :: [P1] Reject malformed xAI usage responses before publishing zero spend | [P2] Anchor Grok’s local Today value to the local scan time
• reviewed 2026-08-20T06:55:01.271Z sha c44ec7a :: needs changes before merge. :: [P2] Cache Grok scans outside menu rendering
• reviewed 2026-08-20T07:32:30.082Z sha c48d4a2 :: needs changes before merge. :: [P2] Read Grok cost history from the cached publication

[Chipagosfinest]

Addressed the ClawSweeper/Codex source-state findings on the current head:

1. xAI /usage failures stay unavailable (missing chart). Successful empty history is confirmed empty.
2. Grok local ~/.grok/sessions tokens publish even when remote billing fails.
3. xAI Today is the UTC day of updatedAt, not the newest point.
4. Grok Today is the current local-day bucket; idle days are not reused.
5. xAI coverage stays a 30-day vendor source even when the dashboard asks for 365.

Focused tests: GrokLocalSessionScannerTests, XAICostUsageMappingTests, XAIProviderTests, GrokXAISpendCatalogTests — 20 passed.

Redacted local Grok proof from this machine after the fix (no identity):

signals_files=6
aggregate_tokens=1242673
models=['grok-4.6']

@clawsweeper re-review

[clawsweeper]

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event exact_review_queue).
Result: when the review finishes, ClawSweeper will create the durable review comment if needed or update the existing comment in place.

Re-review progress:

• State: Complete
• Detail: The durable review result and its route handoff completed.
• Run: https://github.com/openclaw/clawsweeper/actions/runs/32321899378
• Updated: 2026-08-20T01:41:17.821Z

[Chipagosfinest]

@clawsweeper re-review

Exact head c44ec7abc: both durable findings are fixed, make check passes, 23 focused tests pass, and the PR body now includes redacted real Grok scanner → shared catalog output. All four Codex review threads are resolved.

[clawsweeper]

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event exact_review_queue).
Result: when the review finishes, ClawSweeper will create the durable review comment if needed or update the existing comment in place.

[Chipagosfinest]

@clawsweeper re-review

Exact head c48d4a2bb: rebased onto current main and fixed the macOS shard 0 failure from #3086. That PR scoped localizedProviderDetails to DeepSeek/z.ai, so generic OpenRouter details stay canonical English; the merge test still expected API 金鑰. make check passes; 29 focused tests in 5 suites pass.

[clawsweeper]

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event exact_review_queue).
Result: when the review finishes, ClawSweeper will create the durable review comment if needed or update the existing comment in place.

Re-review progress:

• State: Complete
• Detail: The durable review result and its route handoff completed.
• Run: https://github.com/openclaw/clawsweeper/actions/runs/32344821941
• Updated: 2026-08-20T07:37:14.437Z

[Chipagosfinest]

@clawsweeper re-review

Exact head a4f80675a: Peter merged main (84a4ca725 restored generic title/row L() localization). The OpenRouter popup test now matches that restored catalog. make check and PopupLocalizationTests pass on this head.

[clawsweeper]

🦞🧹
ClawSweeper re-review requested.

I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow sweep.yml, event exact_review_queue).
Result: when the review finishes, ClawSweeper will create the durable review comment if needed or update the existing comment in place.