Show pace for OpenCode Go CLI usage windows

[kentoku24]

Summary

codexbar usage --provider opencodego now returns pace for both the primary (5-hour rolling) and secondary (weekly) usage windows — in the text output (Pace: ... lines) and in the JSON pace payload (pace.primary / pace.secondary).

Previously the CLI pace path covered only .session (codex/claude/ollama) and .weekly (codex/claude/opencode/ollama) allowlists, so OpenCode Go got no pace on either lane even though its snapshot already carries both windows with reset times. The GUI already showed pace for the OpenCode Go monthly (tertiary) lane; this change brings the CLI in line via the descriptor's pace capability.

Scope: CLI-facing pace lanes only. The existing OpenCode Go reset-window (GUI Monthly) pace rules are preserved unchanged. No payload schema, fetcher, or docs changes; the Monthly (tertiary) row in the CLI stays unchanged.

Changes

• Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoProviderDescriptor.swift — extend the descriptor's pace capability: keep the existing reset-window / inferred-monthly-duration rules (GUI Monthly pace unchanged) and add primary: .session(maximumMinutes: 300) and secondary: .weekly so the CLI returns pace for the 5-hour and weekly windows (+5/-1).
• Tests/CodexBarTests/CLISnapshotTests.swift — 4 new tests: text pace lines for 5-hour + weekly, JSON payload primary+secondary, primary-only when weekly usage is missing, and the >5h window guard (+101).

Verification

• swift test --filter CLISnapshotTests → 52/52 pass (incl. the 4 new tests)
• swift test --filter ProviderPaceCapabilityTests → 4/4 pass
• swift test --filter MenuBarPaceTextTests → 2/2 pass
• swift build → exit 0; CLI --help smoke test shows the opencodego provider
• SwiftFormat + SwiftLint on the changed files → 0 violations
• Note: the full make test shard run can abort on pre-existing, environment-dependent Claude OAuth keychain test failures unrelated to this change; make check can report pre-existing violations in unrelated uncommitted work-in-progress files. Both are isolated and documented (not touched here).

[clawsweeper]

🦞👀
ClawSweeper picked this up.

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

[clawsweeper]

Codex review: needs real behavior proof before merge. Reviewed August 15, 2026, 8:55 AM ET / 12:55 UTC.

ClawSweeper review

What this changes

The PR enables five-hour primary and weekly secondary pace output for OpenCode Go in the CLI text and JSON renderers, with regression coverage.

Merge readiness

⛔ Blocked until real behavior proof from a real setup is added - 3 items remain

The patch is a focused, correct descriptor-based extension, but it needs real after-fix CLI proof from an OpenCode Go account before merge.

Priority: P3
Reviewed head: a15887deddd90de1be91baada7cdb04c30a217a2

Review scores

Measure	Result	What it means
Overall readiness	🦪 silver shellfish (2/6)	The implementation and focused coverage are solid, but merge-ready confidence is capped until the changed account-backed CLI path is shown working in a real setup.
Proof confidence	🦪 silver shellfish (2/6)	Needs real behavior proof before merge: The PR body reports tests and a build, but provides no redacted after-fix CLI transcript from a real OpenCode Go account. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Verification

Check	Result	Evidence
Real behavior	Needs proof	Needs real behavior proof before merge: The PR body reports tests and a build, but provides no redacted after-fix CLI transcript from a real OpenCode Go account. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
Evidence reviewed	5 items	Current behavior: Current main declares only calendar-month reset-window pace for OpenCode Go, so its five-hour and weekly snapshot slots are not classified for CLI pace. Supported window data: The OpenCode Go snapshot already supplies a 300-minute primary window and a 10,080-minute secondary window, each with a reset time; monthly remains tertiary. Renderer integration: The shared CLI renderer evaluates primary, secondary, and tertiary slots through the descriptor capability, so adding the two lanes is the existing integration point.
Findings	None	None.
Security	None	None.

How this fits together

OpenCode Go fetchers convert account usage into primary, secondary, and monthly usage windows. The provider descriptor classifies those windows for the shared CLI pace renderer, which emits text and JSON usage output.

flowchart LR
    A[OpenCode Go account usage] --> B[Usage snapshot windows]
    B --> C[Provider pace capability]
    C --> D[Shared pace calculation]
    D --> E[CLI text output]
    D --> F[CLI JSON payload]

Loading

Before merge

• Add real behavior proof - Needs real behavior proof before merge: The PR body reports tests and a build, but provides no redacted after-fix CLI transcript from a real OpenCode Go account. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
• Resolve merge risk (P1) - The supplied verification is limited to tests and build claims; without a redacted real CLI run, the account-backed OpenCode Go path remains unverified.
• Complete next step (P2) - A real account-backed CLI run must be supplied by the contributor; there is no safe automated repair that can establish that proof.

Agent review details

Security

None.

Review metrics

Metric	Value	Why it matters
Production versus test delta	production +4 net (+5/-1); tests +101	A small provider-metadata change is backed by focused text, JSON, missing-window, and duration-guard coverage.

Merge-risk options

Maintainer options:

1. Decide the mitigation before merge
   Preserve the descriptor-level monthly rules and land the added primary and secondary lanes after a redacted live codexbar usage --provider opencodego transcript confirms both text and JSON output.
2. Pause or close
   Do not merge this PR until maintainers decide whether the risk is worth taking.

Technical review

Best possible solution:

Preserve the descriptor-level monthly rules and land the added primary and secondary lanes after a redacted live codexbar usage --provider opencodego transcript confirms both text and JSON output.

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

Not applicable: this is a feature extension; the relevant confirmation is an after-fix run against a real OpenCode Go account rather than reproducing an existing failure.

Is this the best way to solve the issue?

Yes. The descriptor is the established shared capability boundary for CLI pace classification, and the branch preserves the existing monthly reset-window behavior while adding only the two supported snapshot lanes.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against f15f142a7787.

Labels

Label changes:

• add P3: This is a low-risk CLI usability enhancement for an optional provider rather than a broken core workflow.
• add rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🐚 platinum hermit.
• add status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body reports tests and a build, but provides no redacted after-fix CLI transcript from a real OpenCode Go account. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Label justifications:

• P3: This is a low-risk CLI usability enhancement for an optional provider rather than a broken core workflow.
• rating: 🦪 silver shellfish: Overall readiness is 🦪 silver shellfish; proof is 🦪 silver shellfish and patch quality is 🐚 platinum hermit.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body reports tests and a build, but provides no redacted after-fix CLI transcript from a real OpenCode Go account. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Evidence

What I checked:

• Current behavior: Current main declares only calendar-month reset-window pace for OpenCode Go, so its five-hour and weekly snapshot slots are not classified for CLI pace. (Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoProviderDescriptor.swift:63, f15f142a7787)
• Supported window data: The OpenCode Go snapshot already supplies a 300-minute primary window and a 10,080-minute secondary window, each with a reset time; monthly remains tertiary. (Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoUsageSnapshot.swift:74, f15f142a7787)
• Renderer integration: The shared CLI renderer evaluates primary, secondary, and tertiary slots through the descriptor capability, so adding the two lanes is the existing integration point. (Sources/CodexBarCLI/CLIRenderer.swift:549, f15f142a7787)
• Patch and coverage: The branch preserves the monthly reset-window rules, adds session and weekly lanes, and adds four focused CLI snapshot cases including a five-hour upper-bound guard. (Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoProviderDescriptor.swift:66, a15887deddd9)
• Feature provenance: Blame traces the current OpenCode Go descriptor baseline to the v0.50.0-era commit; current main still retains the calendar-only capability, so this work is not already implemented. (Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoProviderDescriptor.swift:63, 0e453c4a5b2a)

Likely related people:

• steipete: Current-main blame attributes the OpenCode Go descriptor baseline to Peter Steinberger, and the repository is owned by this account. (role: current descriptor contributor; confidence: medium; commits: 0e453c4a5b2a; files: Sources/CodexBarCore/Providers/OpenCodeGo/OpenCodeGoProviderDescriptor.swift)

Rank-up moves

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

• Add a redacted terminal capture of codexbar usage --provider opencodego showing five-hour and weekly pace in text and JSON, then update the PR body for re-review.

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.

[kentoku24]

Live CLI verification — primary/secondary pace now returned

Built the PR head (a15887ded) in a worktree (swift build -c release --product CodexBarCLI, exit 0) and ran codexbar usage --provider opencodego against the local opencode DB (source: local — no network, no Keychain prompts) with ~/.codexbar/config.json (opencodego enabled). Compared against the installed release (CodexBar 0.49.3).

AFTER (PR head a15887d) — text, exit 0

== OpenCode Go (local) ==
5-hour: 93% left [===========-]
Pace: 93% in reserve | Expected 100% used | Lasts until reset
Resets now
Weekly: 79% left [=========---]
Pace: 72% in reserve | Expected 93% used | Lasts until reset
Resets in 1d 7h
Monthly: 75% left [=========---]
Pace: 17% in reserve | Expected 42% used | Lasts until reset
Resets in 17d 22h

AFTER (PR head a15887d) — JSON pace object, exit 0

"pace" : {
  "primary"   : { "deltaPercent" : -93, "expectedUsedPercent" : 100, "stage" : "farBehind", "summary" : "93% in reserve | Expected 100% used | Lasts until reset", "willLastToReset" : true },
  "secondary" : { "deltaPercent" : -72, "expectedUsedPercent" : 93,  "stage" : "farBehind", "summary" : "72% in reserve | Expected 93% used | Lasts until reset",  "willLastToReset" : true },
  "tertiary"  : { "deltaPercent" : -17, "expectedUsedPercent" : 42,  "stage" : "farBehind", "summary" : "17% in reserve | Expected 42% used | Lasts until reset",  "willLastToReset" : true }
}

BEFORE (installed 0.49.3) — text, exit 0

== OpenCode Go (local) ==
5-hour: 93% left [===========-]
Resets in 1m
Weekly: 79% left [=========---]
Resets in 1d 7h
Monthly: 75% left [=========---]
Pace: 17% in reserve | Expected 42% used | Lasts until reset
Resets in 17d 22h

BEFORE JSON pace contains only tertiary (pace.primary / pace.secondary absent) — reproducing the reported "only tertiary returns pace" behavior.

Verdict

Window	After (a15887d)	Before (0.49.3)
primary (5-hour, 300 min)	Pace ✅ (pace.primary)	no Pace ❌
secondary (Weekly, 10080 min)	Pace ✅ (pace.secondary)	no Pace ❌
tertiary (Monthly, 43200 min)	Pace ✅ (pace.tertiary)	Pace ✅ (pace.tertiary)

Usage numbers are identical before/after (6.8% / 20.8% / 24.9% used); only pace presence differs. No secrets in output (local DB source, no cookies/keys), no Keychain prompts.