一键导入
test-changed
Run e2e and unit tests related to files changed on the current branch. Use after making code changes to verify nothing is broken.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Run e2e and unit tests related to files changed on the current branch. Use after making code changes to verify nothing is broken.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
Commit a logical unit of work with a descriptive message. Invoke after completing a meaningful chunk (component created, store wired, tests passing) to maintain small, reviewable commits.
Guide for building React components using @stellar/design-system and project SCSS conventions. Invoke when creating new components or reviewing component code.
Bridge Figma MCP output to Stellar Lab code conventions. Invoke after calling get_design_context to translate Figma's reference code into project-appropriate components, classes, and patterns.
Correct mental models for React re-renders and memoization. Use this skill when writing, reviewing, or optimizing React components to avoid common misconceptions about performance. Debunks the myth that "props cause re-renders" and teaches when memoization actually helps.
Reference for the two Zustand store patterns in Stellar Lab — main store (querystring) and transaction flow store (sessionStorage). Invoke when creating store slices, debugging hydration, or deciding where state belongs.
基于 SOC 职业分类
| name | test-changed |
| description | Run e2e and unit tests related to files changed on the current branch. Use after making code changes to verify nothing is broken. |
Run e2e and unit tests that are relevant to the files changed on the current branch.
jest.config.js, path aliases via moduleNameMapperplaywright.config.ts, Chromium only, 20s timeout per test, 2 retries, 2 workerspnpm dev automatically, but reuseExistingServer: true in non-CI — a stale server from another branch will cause false failuresKill any stale dev server on port 3000 before running e2e tests. Playwright's reuseExistingServer will pick up a server from a different branch otherwise, causing false failures (especially in worktrees):
lsof -ti:3000 | xargs kill -9 2>/dev/null; true
Find changed files relative to main:
git diff --name-only main...HEAD
git diff --name-only # unstaged
git diff --name-only --cached # staged
Map source changes to test files. Use these rules:
src/app/(sidebar)/transaction/build/** -> tests/e2e/buildTransaction.test.ts, tests/e2e/savedTransactions.test.ts, tests/e2e/urlParams.test.tssrc/app/(sidebar)/transaction/sign/** -> tests/e2e/signTransactionPage.test.tssrc/app/(sidebar)/transaction/submit/** -> tests/e2e/submitTransactionPage.test.tssrc/app/(sidebar)/transaction/simulate/** -> tests/e2e/simulateTransactionPage.test.tssrc/app/(sidebar)/transaction/fee-bump/** -> tests/e2e/feeBumpTransactionPage.test.tssrc/app/(sidebar)/account/fund/** or src/app/(sidebar)/account/create/** -> tests/e2e/fundAccountPage.test.ts, tests/e2e/createAccountPage.test.tssrc/app/(sidebar)/xdr/** -> tests/e2e/viewXdrToJsonPage.test.ts, tests/e2e/jsonToXdrPage.test.tssrc/app/(sidebar)/endpoints/** -> tests/e2e/endpointsPage*.test.tssrc/app/(sidebar)/smart-contracts/** -> tests/e2e/contractExplorer*.test.tssrc/components/**, src/helpers/**, src/hooks/**, src/store/**, src/constants/**, src/types/**, src/validate/**, src/styles/**, src/middleware.tstests/e2e/mock/** -> run ALL e2e tests that import from mock/tests/e2e/*.test.ts changed directly -> run those teststests/unit/** changed -> run unit testssrc/helpers/ or src/validate/ changed, also check tests/unit/ for matching unit testsIf XDR encoding or transaction building logic changed (e.g. ClassicTransactionXdr.tsx, SorobanTransactionXdr.tsx, src/helpers/xdr/), hardcoded hash/XDR values in tests may need updating. Flag this to the user before running tests.
Run the identified tests:
npx playwright test <file1> <file2> ... --retries=0 --reporter=listpnpm test:unit <file>pnpm test:e2e and pnpm test:unitReport results. Show pass/fail counts and list any failures with their error messages. For failures, read the relevant test code and source code to suggest fixes.