Skip to content
Latchkey

CI workflow (vitest-dev/vitest)

The CI workflow from vitest-dev/vitest, explained and optimized by Latchkey.

A

CI health: A - excellent

The optimized version below adds job timeouts.

Source: vitest-dev/vitest.github/workflows/ci.ymlLicense MITView source

What it does

This is the CI workflow from the vitest-dev/vitest repository, a real project running GitHub Actions. It is shown here with attribution under its MIT license.

Below, Latchkey shows a faster, safer version produced by its optimization engine.

The workflow

workflow (.yml)
name: CI

# Remove default permissions of GITHUB_TOKEN for security
# https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobs
permissions: {}

on:
  push:
    branches:
      - main

  pull_request:

  workflow_dispatch:

concurrency:
  group: ci-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

env:
  PLAYWRIGHT_BROWSERS_PATH: ${{ github.workspace }}/.cache/ms-playwright
  # Trust the committed lockfile and skip pnpm's per-install supply-chain
  # verification (trustPolicy: no-downgrade, ~15-20s on every job). The lint job
  # overrides this to false so the lockfile is still verified once per run.
  PNPM_CONFIG_TRUST_LOCKFILE: true

jobs:
  lint:
    timeout-minutes: 10
    runs-on: ubuntu-latest
    name: 'Lint: node-latest, ubuntu-latest'
    # verify the lockfile against supply-chain policies once per run
    env:
      PNPM_CONFIG_TRUST_LOCKFILE: false
    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - uses: ./.github/actions/setup-and-cache

      - name: Install
        run: pnpm i --filter '!docs'

      - name: Build
        run: pnpm run build

      - name: Generate CLI docs
        run: pnpm -C docs run cli-table

      # check uncommited LICENSE.md, auto-imports.d.ts, etc...
      - name: Check stale build artifacts
        run: git diff --exit-code

      - name: Lint
        run: pnpm run lint

      - name: Typecheck
        run: pnpm run typecheck

      - name: Typecheck UI
        run: pnpm run -C packages/ui typecheck:client

      # From https://github.com/rhysd/actionlint/blob/main/docs/usage.md#use-actionlint-on-github-actions
      - name: Check workflow files
        run: |
          bash <(curl https://raw.githubusercontent.com/rhysd/actionlint/main/scripts/download-actionlint.bash)
          ./actionlint -color -shellcheck=""

  changed:
    runs-on: ubuntu-latest
    name: 'Diff: node-latest, ubuntu-latest'
    outputs:
      should_skip: ${{ steps.changed-files.outputs.only_changed == 'true' }}

    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - name: Get changed files
        id: changed-files
        uses: tj-actions/changed-files@22103cc46bda19c2b464ffe86db46df6922fd323 # v47.0.5
        with:
          files: |
            docs/**
            .github/**
            !.github/workflows/ci.yml
            **.md

  test:
    needs: changed
    name: 'Test: ${{ matrix.suite }}, node-${{ matrix.node_version }}, ${{ matrix.os }}'
    if: needs.changed.outputs.should_skip != 'true'
    runs-on: ${{ matrix.os }}

    timeout-minutes: 30

    strategy:
      # Split the heavy e2e and coverage suites onto their own runners so they run
      # in parallel with the unit suite instead of sequentially on a single machine.
      # Linux runs the full matrix on node 24/26; node 22 (the minimum supported
      # version, currently in maintenance LTS) gets a single e2e leg as a floor
      # smoke. macOS runs one e2e leg as an Apple-hardware smoke; unit/coverage are
      # redundant on both.
      matrix:
        suite: [unit, e2e, coverage]
        os: [ubuntu-latest]
        node_version: [24, 26]
        include:
          - suite: e2e
            os: ubuntu-latest
            node_version: 22
          - suite: e2e
            os: macos-latest
            node_version: 24
          - suite: unit
            os: windows-latest
            node_version: 24
          - suite: e2e
            os: windows-latest
            node_version: 24
          - suite: coverage
            os: windows-latest
            node_version: 24
      fail-fast: false

    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - uses: ./.github/actions/setup-and-cache
        with:
          node-version: ${{ matrix.node_version }}

      - name: Install
        run: pnpm i --filter '!docs'

      - uses: ./.github/actions/setup-playwright
        with:
          browsers: chromium

      - name: Build
        run: pnpm run build

      - name: Test
        # matrix.suite is a fixed, workflow-defined value (unit/e2e/coverage)
        run: pnpm run test:ci:${{ matrix.suite }} # zizmor: ignore[template-injection]
        env:
          VITEST_CI_BLOB_LABEL: ${{ matrix.os }}-node-${{ matrix.node_version }}

      - name: Test Examples
        if: matrix.suite == 'unit' && matrix.os == 'ubuntu-latest'
        run: pnpm run test:examples

      - name: Unit Test UI
        if: matrix.suite == 'unit'
        run: pnpm -C packages/ui test:ui
        env:
          VITEST_CI_BLOB_LABEL: ${{ matrix.os }}-node-${{ matrix.node_version }}

      - uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
        with:
          name: vitest-results-${{ matrix.suite }}-${{ matrix.os }}-node-${{ matrix.node_version }}
          path: |
            README.md
            test/unit/.vitest
            test/e2e/.vitest
            packages/ui/.vitest
          retention-days: 1
          include-hidden-files: true

  test-browser:
    needs: changed
    name: 'Browsers: ${{ matrix.name }}, node-24, ${{ matrix.os }}'
    if: needs.changed.outputs.should_skip != 'true'

    runs-on: ${{ matrix.os }}
    strategy:
      # runner.test.ts is the heaviest spec (~28% of the suite, ~70% of whichever
      # shard it lands in). Windows isolates it in its own job and shards the other
      # 32 specs; Ubuntu shards the whole suite 2 ways (runner rides in one shard).
      # Every spec, runner included, runs on both OSes.
      matrix:
        include:
          - os: windows-latest
            name: runner
            command: pnpm run test:browser:playwright runner.test.ts
          - os: windows-latest
            name: rest 1/2
            command: "pnpm run test:browser:playwright --exclude 'specs/runner.test.ts' --shard=1/2"
          - os: windows-latest
            name: rest 2/2
            command: "pnpm run test:browser:playwright --exclude 'specs/runner.test.ts' --shard=2/2"
          - os: ubuntu-latest
            name: shard 1/2
            command: pnpm run test:browser:playwright --shard=1/2
          - os: ubuntu-latest
            name: shard 2/2
            command: pnpm run test:browser:playwright --shard=2/2
      fail-fast: false

    timeout-minutes: 30

    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - uses: ./.github/actions/setup-and-cache
        with:
          node-version: 24

      - name: Install
        run: pnpm i --filter '!docs'

      - uses: ./.github/actions/setup-playwright
        with:
          browsers: all

      - name: Build
        run: pnpm run build

      - name: Test Browser (playwright)
        # matrix.command is a fixed, workflow-defined string
        run: ${{ matrix.command }} # zizmor: ignore[template-injection]

  test-vite7:
    needs: changed
    # Runs on ubuntu to keep macOS under its ~5 concurrent-runner limit.
    name: 'Test: vite@7, ${{ matrix.suite }}, node-24, ubuntu-latest'
    if: needs.changed.outputs.should_skip != 'true'
    runs-on: ubuntu-latest

    timeout-minutes: 30

    strategy:
      # vite7 is a compatibility smoke, not a coverage gate. Run the e2e suite
      # (which exercises real config resolution, transform and runs) and the
      # browser suite on separate runners in parallel.
      matrix:
        suite: [e2e, browser]
      fail-fast: false

    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - uses: ./.github/actions/setup-and-cache
        with:
          node-version: 24

      - name: Install
        run: |
          pnpm override-vite7
          # ubuntu runners have no default git identity; set one for this commit
          git add . && git -c user.email=ci@vitest.dev -c user.name=vitest-ci commit -m "ci" && pnpm i --prefer-offline --no-frozen-lockfile --filter '!docs'

      - uses: ./.github/actions/setup-playwright
        with:
          browsers: all

      - name: Build
        run: pnpm run build

      - name: Test E2E
        if: ${{ matrix.suite == 'e2e' }}
        run: pnpm run test:ci:e2e

      - name: Test Browser (playwright)
        if: ${{ matrix.suite == 'browser' && !cancelled() }}
        # smoke: runner.test.ts exercises the core browser-runner behaviour across
        # every instance, enough to catch vite7 browser-mode regressions
        run: pnpm run test:browser:playwright runner.test.ts

  merge-reports:
    needs:
      - test
      - changed
    if: needs.changed.outputs.should_skip != 'true' && !cancelled()
    runs-on: ubuntu-latest
    name: Merge Reports
    timeout-minutes: 10
    steps:
      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          persist-credentials: false

      - uses: ./.github/actions/setup-and-cache

      - name: Install
        run: pnpm i --filter '!docs'

      - name: Build
        run: pnpm run build

      - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
        with:
          pattern: vitest-results-*
          merge-multiple: true

      - name: Merge reports
        continue-on-error: true
        run: |
          set +e
          pnpm --filter=./packages/ui --no-bail test:ui --merge-reports
          cp ./packages/ui/.vitest/index.html ./packages/ui/.vitest/test-ui.html
          pnpm --filter=./test/unit --filter=./test/e2e --no-bail --sequential test --merge-reports
          cp -rf ./test/unit/.vitest/index.html ./test/unit/.vitest/test-unit.html
          cp -rf ./test/e2e/.vitest/index.html ./test/e2e/.vitest/test-e2e.html
        env:
          VITEST_CI_MERGE_REPORTS: 'true'

      - name: Upload Test UI Report
        id: upload-ui-report
        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
        with:
          path: ./packages/ui/.vitest/test-ui.html
          archive: false
          retention-days: 7

      - name: Upload Test Unit Report
        id: upload-unit-report
        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
        with:
          path: ./test/unit/.vitest/test-unit.html
          archive: false
          retention-days: 7

      - name: Upload Test E2E Report
        id: upload-e2e-report
        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
        with:
          path: ./test/e2e/.vitest/test-e2e.html
          archive: false
          retention-days: 7

      - name: Link CI reports
        run: |
          echo "::notice title=CI reports::test/unit: ${STEPS_UPLOAD_UNIT_REPORT_OUTPUTS_ARTIFACT_URL}"
          echo "::notice title=CI reports::test/e2e: ${STEPS_UPLOAD_E2E_REPORT_OUTPUTS_ARTIFACT_URL}"
          echo "::notice title=CI reports::packages/ui: ${STEPS_UPLOAD_UI_REPORT_OUTPUTS_ARTIFACT_URL}"
        env:
          STEPS_UPLOAD_UNIT_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-unit-report.outputs.artifact-url }}
          STEPS_UPLOAD_E2E_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-e2e-report.outputs.artifact-url }}
          STEPS_UPLOAD_UI_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-ui-report.outputs.artifact-url }}

The same workflow, on Latchkey

Removes redundant runs and caps runaway jobs. Added and changed lines are highlighted.

name: CI # Remove default permissions of GITHUB_TOKEN for security# https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobspermissions: {} on:  push:    branches:      - main   pull_request:   workflow_dispatch: concurrency:  group: ci-${{ github.event.pull_request.number || github.ref }}  cancel-in-progress: true env:  PLAYWRIGHT_BROWSERS_PATH: ${{ github.workspace }}/.cache/ms-playwright  # Trust the committed lockfile and skip pnpm's per-install supply-chain  # verification (trustPolicy: no-downgrade, ~15-20s on every job). The lint job  # overrides this to false so the lockfile is still verified once per run.  PNPM_CONFIG_TRUST_LOCKFILE: true jobs:  lint:    timeout-minutes: 10    runs-on: latchkey-small    name: 'Lint: node-latest, ubuntu-latest'    # verify the lockfile against supply-chain policies once per run    env:      PNPM_CONFIG_TRUST_LOCKFILE: false    steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - uses: ./.github/actions/setup-and-cache       - name: Install        run: pnpm i --filter '!docs'       - name: Build        run: pnpm run build       - name: Generate CLI docs        run: pnpm -C docs run cli-table       # check uncommited LICENSE.md, auto-imports.d.ts, etc...      - name: Check stale build artifacts        run: git diff --exit-code       - name: Lint        run: pnpm run lint       - name: Typecheck        run: pnpm run typecheck       - name: Typecheck UI        run: pnpm run -C packages/ui typecheck:client       # From https://github.com/rhysd/actionlint/blob/main/docs/usage.md#use-actionlint-on-github-actions      - name: Check workflow files        run: |          bash <(curl https://raw.githubusercontent.com/rhysd/actionlint/main/scripts/download-actionlint.bash)          ./actionlint -color -shellcheck=""   changed:    timeout-minutes: 30    runs-on: latchkey-small    name: 'Diff: node-latest, ubuntu-latest'    outputs:      should_skip: ${{ steps.changed-files.outputs.only_changed == 'true' }}     steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - name: Get changed files        id: changed-files        uses: tj-actions/changed-files@22103cc46bda19c2b464ffe86db46df6922fd323 # v47.0.5        with:          files: |            docs/**            .github/**            !.github/workflows/ci.yml            **.md   test:    needs: changed    name: 'Test: ${{ matrix.suite }}, node-${{ matrix.node_version }}, ${{ matrix.os }}'    if: needs.changed.outputs.should_skip != 'true'    runs-on: latchkey-small     timeout-minutes: 30     strategy:      # Split the heavy e2e and coverage suites onto their own runners so they run      # in parallel with the unit suite instead of sequentially on a single machine.      # Linux runs the full matrix on node 24/26; node 22 (the minimum supported      # version, currently in maintenance LTS) gets a single e2e leg as a floor      # smoke. macOS runs one e2e leg as an Apple-hardware smoke; unit/coverage are      # redundant on both.      matrix:        suite: [unit, e2e, coverage]        os: [ubuntu-latest]        node_version: [24, 26]        include:          - suite: e2e            os: ubuntu-latest            node_version: 22          - suite: e2e            os: macos-latest            node_version: 24          - suite: unit            os: windows-latest            node_version: 24          - suite: e2e            os: windows-latest            node_version: 24          - suite: coverage            os: windows-latest            node_version: 24      fail-fast: false     steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - uses: ./.github/actions/setup-and-cache        with:          node-version: ${{ matrix.node_version }}       - name: Install        run: pnpm i --filter '!docs'       - uses: ./.github/actions/setup-playwright        with:          browsers: chromium       - name: Build        run: pnpm run build       - name: Test        # matrix.suite is a fixed, workflow-defined value (unit/e2e/coverage)        run: pnpm run test:ci:${{ matrix.suite }} # zizmor: ignore[template-injection]        env:          VITEST_CI_BLOB_LABEL: ${{ matrix.os }}-node-${{ matrix.node_version }}       - name: Test Examples        if: matrix.suite == 'unit' && matrix.os == 'ubuntu-latest'        run: pnpm run test:examples       - name: Unit Test UI        if: matrix.suite == 'unit'        run: pnpm -C packages/ui test:ui        env:          VITEST_CI_BLOB_LABEL: ${{ matrix.os }}-node-${{ matrix.node_version }}       - uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1        with:          name: vitest-results-${{ matrix.suite }}-${{ matrix.os }}-node-${{ matrix.node_version }}          path: |            README.md            test/unit/.vitest            test/e2e/.vitest            packages/ui/.vitest          retention-days: 1          include-hidden-files: true   test-browser:    needs: changed    name: 'Browsers: ${{ matrix.name }}, node-24, ${{ matrix.os }}'    if: needs.changed.outputs.should_skip != 'true'     runs-on: latchkey-small    strategy:      # runner.test.ts is the heaviest spec (~28% of the suite, ~70% of whichever      # shard it lands in). Windows isolates it in its own job and shards the other      # 32 specs; Ubuntu shards the whole suite 2 ways (runner rides in one shard).      # Every spec, runner included, runs on both OSes.      matrix:        include:          - os: windows-latest            name: runner            command: pnpm run test:browser:playwright runner.test.ts          - os: windows-latest            name: rest 1/2            command: "pnpm run test:browser:playwright --exclude 'specs/runner.test.ts' --shard=1/2"          - os: windows-latest            name: rest 2/2            command: "pnpm run test:browser:playwright --exclude 'specs/runner.test.ts' --shard=2/2"          - os: ubuntu-latest            name: shard 1/2            command: pnpm run test:browser:playwright --shard=1/2          - os: ubuntu-latest            name: shard 2/2            command: pnpm run test:browser:playwright --shard=2/2      fail-fast: false     timeout-minutes: 30     steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - uses: ./.github/actions/setup-and-cache        with:          node-version: 24       - name: Install        run: pnpm i --filter '!docs'       - uses: ./.github/actions/setup-playwright        with:          browsers: all       - name: Build        run: pnpm run build       - name: Test Browser (playwright)        # matrix.command is a fixed, workflow-defined string        run: ${{ matrix.command }} # zizmor: ignore[template-injection]   test-vite7:    needs: changed    # Runs on ubuntu to keep macOS under its ~5 concurrent-runner limit.    name: 'Test: vite@7, ${{ matrix.suite }}, node-24, ubuntu-latest'    if: needs.changed.outputs.should_skip != 'true'    runs-on: latchkey-small     timeout-minutes: 30     strategy:      # vite7 is a compatibility smoke, not a coverage gate. Run the e2e suite      # (which exercises real config resolution, transform and runs) and the      # browser suite on separate runners in parallel.      matrix:        suite: [e2e, browser]      fail-fast: false     steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - uses: ./.github/actions/setup-and-cache        with:          node-version: 24       - name: Install        run: |          pnpm override-vite7          # ubuntu runners have no default git identity; set one for this commit          git add . && git -c user.email=ci@vitest.dev -c user.name=vitest-ci commit -m "ci" && pnpm i --prefer-offline --no-frozen-lockfile --filter '!docs'       - uses: ./.github/actions/setup-playwright        with:          browsers: all       - name: Build        run: pnpm run build       - name: Test E2E        if: ${{ matrix.suite == 'e2e' }}        run: pnpm run test:ci:e2e       - name: Test Browser (playwright)        if: ${{ matrix.suite == 'browser' && !cancelled() }}        # smoke: runner.test.ts exercises the core browser-runner behaviour across        # every instance, enough to catch vite7 browser-mode regressions        run: pnpm run test:browser:playwright runner.test.ts   merge-reports:    needs:      - test      - changed    if: needs.changed.outputs.should_skip != 'true' && !cancelled()    runs-on: latchkey-small    name: Merge Reports    timeout-minutes: 10    steps:      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2        with:          persist-credentials: false       - uses: ./.github/actions/setup-and-cache       - name: Install        run: pnpm i --filter '!docs'       - name: Build        run: pnpm run build       - uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0        with:          pattern: vitest-results-*          merge-multiple: true       - name: Merge reports        continue-on-error: true        run: |          set +e          pnpm --filter=./packages/ui --no-bail test:ui --merge-reports          cp ./packages/ui/.vitest/index.html ./packages/ui/.vitest/test-ui.html          pnpm --filter=./test/unit --filter=./test/e2e --no-bail --sequential test --merge-reports          cp -rf ./test/unit/.vitest/index.html ./test/unit/.vitest/test-unit.html          cp -rf ./test/e2e/.vitest/index.html ./test/e2e/.vitest/test-e2e.html        env:          VITEST_CI_MERGE_REPORTS: 'true'       - name: Upload Test UI Report        id: upload-ui-report        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1        with:          path: ./packages/ui/.vitest/test-ui.html          archive: false          retention-days: 7       - name: Upload Test Unit Report        id: upload-unit-report        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1        with:          path: ./test/unit/.vitest/test-unit.html          archive: false          retention-days: 7       - name: Upload Test E2E Report        id: upload-e2e-report        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1        with:          path: ./test/e2e/.vitest/test-e2e.html          archive: false          retention-days: 7       - name: Link CI reports        run: |          echo "::notice title=CI reports::test/unit: ${STEPS_UPLOAD_UNIT_REPORT_OUTPUTS_ARTIFACT_URL}"          echo "::notice title=CI reports::test/e2e: ${STEPS_UPLOAD_E2E_REPORT_OUTPUTS_ARTIFACT_URL}"          echo "::notice title=CI reports::packages/ui: ${STEPS_UPLOAD_UI_REPORT_OUTPUTS_ARTIFACT_URL}"        env:          STEPS_UPLOAD_UNIT_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-unit-report.outputs.artifact-url }}          STEPS_UPLOAD_E2E_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-e2e-report.outputs.artifact-url }}          STEPS_UPLOAD_UI_REPORT_OUTPUTS_ARTIFACT_URL: ${{ steps.upload-ui-report.outputs.artifact-url }} 

What changed

  • Run on Latchkey managed runners with one line (runs-on), which apply the fixes below automatically and self-heal transient failures. This example uses latchkey-small; pick the runner size that fits the job.
  • Add a job timeout so a hung step cannot burn hours of runner time.

What Latchkey heals here

This workflow has steps that commonly fail on transient issues (network, registries, flaky browsers). On Latchkey managed runners they are detected, retried, and self-healed instead of failing your build:

  • End-to-end and browser tests

This workflow runs 6 jobs (12 with the matrix expanded) per trigger. On Latchkey the same minutes cost up to 58% less than GitHub-hosted, with zero queue time.

Actions used in this workflow

Frequently asked questions

What does the CI workflow (vitest-dev/vitest) workflow do?
This is the CI workflow from the vitest-dev/vitest repository, a real project running GitHub Actions. It is shown here with attribution under its MIT license.
What CI health grade does this workflow get?
This Automation and other workflow grades A. Paste your own workflow into the Latchkey grader to see its grade and the exact fixes.
How can I improve this Automation and other workflow?
Apply job timeouts. Latchkey applies these automatically on managed runners when you point runs-on at Latchkey.

References