no API key found — this repo is configured to use deepseek/deepseek-v4-pro, which needs DEEPSEEK_API_KEY, but the runner has no key for it.

To fix: add the key as a GitHub Actions secret (referenced from your workflow's env: block) or as a Pullfrog secret in the console — or switch this repo to a different model (free models need no key).

Open repo secrets → · Configure model → · Setup docs → · Ask in Discord →

  ^{｜ Rerun failed job ➔ ｜ View workflow run ｜ via Pullfrog ｜ Using DeepSeek Pro (free via Pullfrog for OSS) ｜ 𝕏}

@coderabbitai review
@cubic-dev-ai review

@coderabbitai review
@cubic-dev-ai review

@ryoppippi I have started the AI code review. It will take a few minutes to complete.

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

#!/bin/bash
# Fetch PR details and diff
gh pr view 1306 --repo ccusage/ccusage --json title,body,files,additions,deletions,changedFiles

echo "---FILES---"
gh pr diff 1306 --repo ccusage/ccusage

@ryoppippi: I've reviewed the changes in this PR. Here's my assessment:

Summary

The Nushell port is a clean, faithful translation of the original Bun/TypeScript script. All core paths are preserved — native package fast path with portable-binary validation, version check, and cargo fallback build.

Observations

✅ Logic is faithfully preserved

✅ Idiomatic Nushell

• $env.CURRENT_FILE | path dirname | path join ../../.. | path expand is the correct way to locate the repo root from a script path.
• get --optional $target_key for safe record access is idiomatic.
• run-external $binary '--version' | complete is the correct form when the binary path is in a variable (the ^var short-form doesn't support arbitrary path variables in all Nushell versions).
• Short-circuit and in native_package_includes_binary is safe — Nushell does evaluate and/or lazily.

🔍 Minor observations

is_portable_binary — ldd coverage on Linux ARM64

ldd on musl-compiled static binaries correctly prints statically linked; on some ARM64 distributions it may print not a dynamic executable instead. The regex covers both cases, which is consistent with the original.

cargo_binary path never copied to native package dir

This is consistent with the TypeScript version — the cargo fallback is used only in dev/local scenarios; CI pre-builds and downloads artifacts. No change needed, just confirming parity.

nix/dev-shell.nix — nushell added alongside bun

This is correct per the PR description: other TS scripts (bench, large-fixture) still use Bun APIs.

Verdict

The migration is well-executed and the logic is equivalent to the original. CI/release workflow comments are updated to reflect the change. No functional issues found.

Open in StackBlitz

npx https://pkg.pr.new/ccusage@1306

npx https://pkg.pr.new/@ccusage/ccusage-darwin-arm64@1306

npx https://pkg.pr.new/@ccusage/ccusage-darwin-x64@1306

npx https://pkg.pr.new/@ccusage/ccusage-linux-arm64@1306

npx https://pkg.pr.new/@ccusage/ccusage-linux-x64@1306

npx https://pkg.pr.new/@ccusage/ccusage-win32-x64@1306

commit: 107ca57

ccusage performance comparison

PR SHA: 98c3abe1ac04
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: 98c3abe. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

@cubic-dev-ai addressed your P3 drift-risk finding in 82b37b2 by deriving the native package root from the package manifests os/cpu metadata instead of keeping a separate hardcoded map in the Nushell script.

Validation after the fix:

• nix develop --command pnpm --filter=./apps/ccusage run ensure:native-binary
• fallback path with packages/ccusage-darwin-arm64/bin/ccusage temporarily moved aside, then restored
• nix develop --command just fmt
• nix develop --command just ccusage::build

@coderabbitai review

ccusage performance comparison

PR SHA: 98c3abe1ac04
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: 98c3abe. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

@cubic-dev-ai addressed your P3 drift-risk finding in 82b37b2 by deriving the native package root from the package manifests os/cpu metadata instead of keeping a separate hardcoded map in the Nushell script.

Validation after the fix:

• nix develop --command pnpm --filter=./apps/ccusage run ensure:native-binary
• fallback path with packages/ccusage-darwin-arm64/bin/ccusage temporarily moved aside, then restored
  ...

@ryoppippi I have started the AI code review. It will take a few minutes to complete.

#!/bin/bash
# Get the latest diff of the PR
gh pr diff 1306 --repo ccusage/ccusage

@ryoppippi: The fix in 82b37b28 looks great. The new matching_native_package_root function is a clear improvement:

glob ($repo_root | path join packages 'ccusage-*' package.json)
| each {|package_json_path|
    let os  = $package_json.os?
    let cpu = $package_json.cpu?
    if (list_contains $os $target_platform) and (list_contains $cpu $target_arch) { ... }
}
| where {|package_root| $package_root != null }

A few things I verified:

• $package_json.os? / .cpu? — optional field access is correct; returns null if absent, which list_contains handles safely via the describe type-guard.
• ($value | describe) =~ '^list' — Nushell's describe returns list<string> (or similar) for list values, so the '^list' regex is the right guard.
• $candidates | get --optional 0 — correctly returns null with no match rather than erroring.
• Drift-risk eliminated — adding a new platform package now requires no script change; the script auto-discovers it from the manifest os/cpu fields.

The rest of the script is unchanged from the previous review. Everything looks good. ✅

[approve]

ccusage performance comparison

PR SHA: 82b37b289dd5
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: 82b37b2. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: 82b37b289dd5
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: 82b37b2. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

@coderabbitai review
@cubic-dev-ai review

Pushed c4545e1 to use a Nix shebang for the Nushell script and remove nushell from the lightweight CI/release profile installs. I also checked the remaining small scripts and documented why stage-native-package.mjs should stay JS for the Windows action path.

@coderabbitai review
@cubic-dev-ai review

Pushed c4545e1 to use a Nix shebang for the Nushell script and remove nushell from the lightweight CI/release profile installs. I also checked the remaining small scripts and documented why stage-native-package.mjs should stay JS for the Windows action path.

@ryoppippi I have started the AI code review. It will take a few minutes to complete.

#!/bin/bash
gh pr diff 1306 --repo ccusage/ccusage

@ryoppippi: Commit c4545e17 looks great. Here's what I verified:

✅ Nix shebang — correct form and path

#!/usr/bin/env nix
#! nix shell --inputs-from ../../.. nixpkgs#nushell --command nu

The multi-line Nix shebang is the standard form. The --inputs-from ../../.. path is resolved by Nix relative to the script file's location (apps/ccusage/scripts/), so three levels up lands at the repo root — exactly matching the $env.CURRENT_FILE | path dirname | path join ../../.. | path expand logic used in the script body. Consistent and correct.

✅ CI/release profile installs trimmed correctly

Nushell is no longer an explicit profile install — it's self-provisioned on first invocation through the shebang. nodejs is retained in CI (needed for tsdown); it was never in the release job because setup-node handles it there.

✅ Direct invocation in package.json

scripts/ensure-native-binary.nu (no nu prefix) is correct because the file is marked executable (mode 100755) and the shebang handles runtime dispatch. Script logic is unchanged from the previously reviewed commit.

Everything is clean. The self-bootstrapping pattern is a nice quality-of-life improvement — one fewer thing to keep in sync across ci.yaml, release.yaml, and dev-shell.nix. ✅

[approve]

ccusage performance comparison

PR SHA: c4545e1793a8
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: c4545e1. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: c4545e1793a8
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: c4545e1. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

@coderabbitai review
@cubic-dev-ai review

Pushed bbe813b to replace the Bun-based performance comment upsert helper with a Nushell script that uses gh api through a Nix shebang. I also checked Nushell markdown support: to md exists, but compare-pr-performance.ts still owns too much measurement/data-processing logic to move in this PR without adding churn.

@coderabbitai review
@cubic-dev-ai review

Pushed bbe813b to replace the Bun-based performance comment upsert helper with a Nushell script that uses gh api through a Nix shebang. I also checked Nushell markdown support: to md exists, but compare-pr-performance.ts still owns too much measurement/data-processing logic to move in this PR without adding churn.

@ryoppippi I have started the AI code review. It will take a few minutes to complete.

#!/bin/bash
gh pr diff 1306 --repo ccusage/ccusage

@ryoppippi: Commit bbe813b0 looks great. Here's what I verified:

✅ Nix shebang — correct path

#!/usr/bin/env nix
#! nix shell --inputs-from ../.. nixpkgs#nushell nixpkgs#gh --command nu

--inputs-from ../.. is relative to .github/scripts/, so two levels up lands at the repo root. Both nushell and gh are provisioned by the shebang — no external install step needed. ✅

✅ Behavior parity with the TypeScript version

✅ CI invocation updated

pnpm exec bun .github/scripts/upsert-pr-comment.ts → .github/scripts/upsert-pr-comment.nu (direct executable invocation). The shebang handles runtime dispatch. ✅

Everything is clean. The switch to gh api removes the need for manual token header management and is idiomatic for CI scripts. ✅

[approve]

ccusage performance comparison

PR SHA: bbe813b0d0ba
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: bbe813b. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs rust/target/release/ccusage directly. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: bbe813b0d0ba
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base package: 5f962cd49e24; PR package: bbe813b. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2,597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2,597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

You're iterating quickly on this pull request. To help protect your rate limits, cubic has paused automatic reviews on new pushes for now—when you're ready for another review, comment @cubic-dev-ai review.

ccusage performance comparison

PR SHA: 973bd9c9ac85
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base package: 5f962cd49e24; PR package: 973bd9c. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: 973bd9c9ac85
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runner startup

Execution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one bunx -p <url> ccusage --version run with an empty Bun install cache. Warm reuses that cache and reports the median of repeated runs.

Cached bunx execution performance

Runs the same large fixture through bunx -p <pkg.pr.new URL> ccusage after the Bun install cache has already been populated by the startup measurement. This separates cached package-runner execution from first-fetch package materialization.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base package: 5f962cd49e24; PR package: 973bd9c. Both run through bunx -p <pkg.pr.new URL> ccusage using the warmed Bun install cache from package runner startup, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: 2a3b1e5f2d14
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

ccusage performance comparison

PR SHA: 2a3b1e5f2d14
Base SHA: 5f962cd49e24

This compares the PR package against the configured base package on the same CI runner.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.

@coderabbitai review
@cubic-dev-ai review

@coderabbitai review
@cubic-dev-ai review

@ryoppippi I have started the AI code review. It will take a few minutes to complete.

@ryoppippi I’ll review the latest changes in this PR.

ccusage performance comparison

PR SHA: b6e9fa9f2d5b
Base SHA: 5f962cd49e24

This compares the Rust PR release binary against the configured base package on the same CI runner.

Package runtime diagnostics

Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.

Committed fixture performance

Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.

Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Large real-world-shaped fixture performance

Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.

Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.

Artifact size

Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.