Skip to main content
skill CX Codex

compare-test-case — Claude Code skill: what it does, how to install

compare-test-case compares one test case across two branches: pass/fail, duration, flakiness and failure details. Useful when a feature branch seems to have broken a test.

no license 255 installs ★ 35 updated 2026-03-27

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add tuist/agent-skills --skill compare-test-case -a claude-code
Open on GitHub

What compare-test-case does

Hand it a test case and two branches, and it lines up how that case behaves on each. The case record comes from tuist test case show <id-or-identifier> --json, carrying reliability, flakiness share and run counts. Runs are fetched next and split by the git_branch field.

Side by side you get pass rate, average duration and how many runs went flaky. The verdict falls into a bucket: newly failing, newly flaky, duration regression past a 50% rise, fixed, or unchanged. Failing runs get opened with tuist test case run show <run-id> --json for the assertion message, source path, line number, retry pattern and any crash report. What comes back is a summary: what changed, a root cause hypothesis, and fixes tied to files and lines.

No case in hand? The skill can start from tuist test case list --flaky --json --page-size 10. Base branch defaults to the project default, usually main, and head is whatever branch is checked out. All of it rides on tuist CLI: with no Tuist project behind it, the first command returns nothing and there is nothing to compare.

What it helps with

  • See what a feature branch broke in one specific test
  • Tell a real regression from a one-off flake
  • Get the file and line behind a failing assertion
  • Check whether a change slowed the test down
  • Pick a flaky test to dig into when none is named

Example prompts for the agent

  • Compare AuthModuleTests/LoginTests/test_login_with_expired_token on main and my current branch, then tell me what changed.
  • Pull the latest runs of test_login_with_expired_token and give me pass rate and average duration per branch.
  • Find the flakiest tests with tuist test case list --flaky and walk through the worst one.

How to install compare-test-case

1Skills CLI (Vercel)

npx skills add tuist/agent-skills --skill compare-test-case -a claude-code

2Manually: put the files in place

~/.claude/skills/compare-test-case/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/compare-test-case/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about compare-test-case

How to install compare-test-case in Claude Code?

Run npx skills add tuist/agent-skills --skill compare-test-case -a claude-code; the skill then sits at ~/.claude/skills/compare-test-case/SKILL.md.

What does compare-test-case do?

It gathers runs of a single test case on two branches and compares pass rate, average duration and flaky runs. The output is a verdict plus a root cause guess with suggested fixes by file and line.

What does it need to work?

tuist CLI and tests that already report into Tuist. Without explicit branches, the project default (usually main) is the base and the checked-out git branch is the head.

Similar skills

All skills for Testing and QA →

From the same set12

The whole set →
migrating-to-tuist-generated-projects migrating-to-tuist-generated-projects is a Claude Code skill that moves an existing Xcode project into Tuist-generated files. It runs during a Tuist adoption and checks that the app builds and launches. CX Codex +1 521installs compare-builds compare-builds analyzes two Xcode runs to spot build duration regressions, cache hit rate drops, and new compile issues via Tuist. CX Codex +1 317installs analyze-selective-testing analyze-selective-testing diagnoses Tuist selective test runs in Xcode to spot skipped targets and identify cache invalidation causes. CX Codex +1 310installs compare-generations compare-generations compares two Tuist generation runs to pinpoint cache hit drops and uncover which modules caused invalidation cascades. CX Codex +1 300installs compare-bundles compare-bundles evaluates two application packages via Tuist to spot size growth and heavy files. Helpful for catching build regressions early. CX Codex +1 296installs compare-cache-runs compare-cache-runs is a Claude Code skill that compares two Tuist cache runs, revealing hit-rate changes, root causes, and cascade effects through dependencies. CX Codex +1 291installs compare-test-runs compare-test-runs compares two Tuist test executions to spot fresh failures, newly flaky suites, and execution duration spikes. CX Codex +1 265installs compare-gradle-builds compare-gradle-builds compares two Gradle build runs to spot duration regressions, cache changes, and task failures. Works with run IDs, dashboard links, or branch names. CX Codex +1 239installs

Project sponsors

Companies that keep this analytics open