Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guidelines for preparing changelog entries for nextest releases following Keep a Changelog format
.claude/skills/microck-prepare-changelog/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 245% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 154% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 137% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 373% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 356% | 0% |
This document describes the format and conventions used in site/src/changelog.md.
The changelog follows the Keep a Changelog format with nextest-specific conventions.
markdown## [X.Y.Z] - YYYY-MM-DD
Sections should appear in this order (only include sections that are relevant):
### for section headers (e.g., ### Added)-Good:
markdown- Nextest can now update itself! Once this version is installed, simply run `cargo nextest self update` to update to the latest version.
Good (with note to distributors):
markdown- Nextest now sets `NEXTEST_LD_*` and `NEXTEST_DYLD_*` environment variables to work around macOS System Integrity Protection sanitization. > Note to distributors: ...
Good (with forward-looking context):
markdown- A new `threads-required` configuration that can be specified as a per-test override. This can be used to limit concurrency for heavier tests, to avoid overwhelming CPU or running out of memory.
([#2618])[#2618]: https://github.com/nextest-rs/nextest/pull/2618/pull/ URL/issues/ URL[text](URL)[GHSA-xxxx](https://github.com/advisories/GHSA-xxxx), [CVE-xxxx](https://nvd.nist.gov/vuln/detail/CVE-xxxx)Always thank first-time contributors using this format (use GitHub username only, not full name):
markdownThanks [username](https://github.com/username) for your first contribution!
Place the attribution:
Examples:
markdown- New feature that does something. Thanks [alice](https://github.com/alice) for your first contribution!
markdown### Added - Feature A - Feature B Thanks [bob](https://github.com/bob) for your first contribution!
For contributors who have contributed before, you can optionally thank them but don't say "first contribution":
markdownThanks [charlie](https://github.com/charlie) for your contribution!
Or simply:
markdownThanks [charlie](https://github.com/charlie)!
When multiple people contributed to a feature:
markdownThanks [alice](https://github.com/alice) and [bob](https://github.com/bob) for your contributions!
Use blockquotes for notes to distributors or package maintainers:
markdown> Note to distributors: you can disable self-update by building cargo-nextest with `--no-default-features`.
For warning about future behavior changes:
markdown### Upcoming behavior changes If no tests are run, nextest will start exiting with the advisory code **4** in versions released after 2024-11-18. See [discussion #1646](https://github.com/nextest-rs/nextest/discussions/1646) for more.
Clearly mark experimental features:
markdown- Experimental support for [feature name](link). Please try them out, and provide feedback in the [tracking issue](link)!
If a release contains breaking changes, consider adding a note at the top:
markdownThis is a major release with several new features. It's gone through a period of beta testing, but if you run into issues please [file a bug]!
cargo nextest run toml , bash When showing configuration:
markdownFor example, to time out after 120 seconds: ```toml slow-timeout = { period = "60s", terminate-after = 2 } ```
Note the indentation for the code block within a bullet point.
NEXTEST_RETRIES NAME=value when showing how to set themList major dependency updates or security updates separately:
markdown- Update rust-openssl for [CVE-2025-24898](https://nvd.nist.gov/vuln/detail/CVE-2025-24898).
markdown### Added - A new `--hide-progress-bar` option (environment variable `NEXTEST_HIDE_PROGRESS_BAR`) forces the progress bar to be hidden. Thanks [Remo Senekowitsch](https://github.com/remlse) for your first contribution!
markdown### Added - Nextest now supports assigning [test priorities](https://nexte.st/docs/configuration/test-priorities) via configuration.
markdown### Fixed - Fixed an occasional hang on Linux with [libtest JSON output](https://nexte.st/docs/machine-readable/libtest-json/). For more details, see [#2316]. [#2316]: https://github.com/nextest-rs/nextest/pull/2316
markdown### Changed - If nextest is unable to parse `--target` (and in particular, a custom target), it now fails rather than printing a warning and assuming the host platform. This is being treated as a bugfix because the previous behavior was incorrect.
To generate a changelog entry:
git log <previous-tag>..main --onelinegit log --all --author="Name" --oneline | wc -lgh pr view <number> --json author --jq '.author.login'git show <commit> --statFilter out:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 5,280 | 7,288 | +38% | 1 | 1 | 0% | 1,052 | 3,625 | +245% | 0 | 0 | — |
case-02 | fail→pass | 6,759 | 5,635 | -17% | 1 | 1 | 0% | 1,276 | 3,245 | +154% | 0 | 0 | — |
case-03 | fail→pass | 7,627 | 6,722 | -12% | 1 | 1 | 0% | 1,428 | 3,379 | +137% | 0 | 0 | — |
case-04 | pass→pass | 4,514 | 4,167 | -8% | 1 | 1 | 0% | 667 | 2,659 | +299% | 0 | 0 | — |
case-05 | pass→pass | 8,450 | 6,196 | -27% | 1 | 1 | 0% | 1,477 | 3,064 | +107% | 0 | 0 | — |
case-06 | pass→pass | 7,958 | 7,040 | -12% | 1 | 1 | 0% | 1,306 | 2,888 | +121% | 0 | 0 | — |
case-07 | pass→pass | 5,871 | 3,041 | -48% | 1 | 1 | 0% | 1,063 | 2,640 | +148% | 0 | 0 | — |
case-08 | fail→pass | 4,727 | 7,380 | +56% | 1 | 1 | 0% | 744 | 3,520 | +373% | 0 | 0 | — |
case-09 | fail→fail | 4,303 | 6,700 | +56% | 1 | 1 | 0% | 731 | 3,167 | +333% | 0 | 0 | — |
case-10 | fail→pass | 3,624 | 3,633 | +0% | 1 | 1 | 0% | 590 | 2,692 | +356% | 0 | 0 | — |
case-11 | pass→pass | 7,465 | 3,882 | -48% | 1 | 1 | 0% | 1,132 | 2,658 | +135% | 0 | 0 | — |
case-12 | pass→pass | 7,383 | 3,713 | -50% | 1 | 1 | 0% | 1,345 | 2,729 | +103% | 0 | 0 | — |
case-13 | fail→pass | 6,628 | 4,403 | -34% | 1 | 1 | 0% | 1,137 | 2,829 | +149% | 0 | 0 | — |
case-14 | fail→fail | 7,670 | 3,434 | -55% | 1 | 1 | 0% | 1,133 | 2,621 | +131% | 0 | 0 | — |
case-15 | fail→pass | 9,765 | 3,568 | -63% | 1 | 1 | 0% | 1,515 | 2,593 | +71% | 0 | 0 | — |
case-16 | fail→pass | 6,667 | 4,484 | -33% | 1 | 1 | 0% | 923 | 2,698 | +192% | 0 | 0 | — |
case-17 | pass→pass | 3,983 | 3,545 | -11% | 1 | 1 | 0% | 626 | 2,636 | +321% | 0 | 0 | — |
case-22 | fail→pass | 7,945 | 3,560 | -55% | 1 | 1 | 0% | 1,410 | 2,635 | +87% | 0 | 0 | — |
case-18 | pass→pass | 4,921 | 3,695 | -25% | 1 | 1 | 0% | 730 | 2,664 | +265% | 0 | 0 | — |
case-19 | pass→pass | 7,666 | 4,203 | -45% | 1 | 1 | 0% | 1,354 | 2,772 | +105% | 0 | 0 | — |
case-20 | pass→pass | 5,125 | 2,438 | -52% | 1 | 1 | 0% | 820 | 2,517 | +207% | 0 | 0 | — |
case-21 | pass→pass | 5,631 | 3,976 | -29% | 1 | 1 | 0% | 821 | 2,662 | +224% | 0 | 0 | — |
case-23 | fail→pass | 5,520 | 1,659 | -70% | 1 | 1 | 0% | 872 | 2,312 | +165% | 0 | 0 | — |
case-24 | fail→pass | 6,836 | 2,324 | -66% | 1 | 1 | 0% | 1,265 | 2,295 | +81% | 0 | 0 | — |
case-25 | pass→pass | 8,682 | 3,249 | -63% | 1 | 1 | 0% | 1,389 | 2,531 | +82% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 25 cases were attempted. The headline lift of +44 percentage points is the difference between those two pass rates over the 25 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.