@a11y-skills/audit 0.4.0 → 0.4.1

package/CHANGELOG.md

@@ -3,6 +3,16 @@
 All notable changes to `@a11y-skills/audit` are documented here. This project
 adheres to [Semantic Versioning](https://semver.org/).
 
+## 0.4.1 — 2026-06-12
+
+### Fixed
+
+- CLI: `orientation-check` and `time-limit-detector` were failing with
+  "No target URL provided" because the registry entries did not pass `targetUrl`
+  to the underlying check functions. Both checks own their own navigation, so the
+  CLI's redundant pre-navigation step is now skipped for these two entries via the
+  new `ownsNavigation: true` flag on `CheckEntry`.
+
 ## 0.4.0 — 2026-06-12
 
 ### Added

package/README.ja.md

@@ -114,6 +114,57 @@ exit コードが非 0 でも、完了したチェックの JSON は必ず書き
 未インストールの場合はそのチェックだけ `SKIPPED` となり、終了コードには影響しません。
 他の 9 検査は通常通り実行されます。
 
+### GitHub Actions
+
+手動トリガーで監査を実行するコピー&ペースト用ワークフロー。violations があるとジョブは fail する（終了コード 1）。結果を情報提供のみにしたい場合は監査ステップに `continue-on-error: true` を付ける。
+
+```yaml
+# .github/workflows/a11y-audit.yml
+name: A11y Audit
+
+on:
+  workflow_dispatch:
+    inputs:
+      target_url:
+        description: 'Audit target URL'
+        required: true
+        default: 'https://example.com'
+
+jobs:
+  a11y-audit:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Setup Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: '24'
+
+      - name: Install audit package
+        run: npm install --no-save @a11y-skills/audit @playwright/test @axe-core/playwright
+
+      - name: Install Playwright browsers
+        run: npx playwright install --with-deps chromium
+
+      - name: Run a11y audit
+        run: npx a11y-audit --url "$TARGET_URL" --output-dir a11y-audit-results
+        env:
+          TARGET_URL: ${{ inputs.target_url }}
+
+      - name: Upload audit results
+        uses: actions/upload-artifact@v7
+        if: always()
+        with:
+          name: a11y-audit-results
+          path: a11y-audit-results/
+          retention-days: 30
+```
+
+Notes:
+
+- `push` / `schedule` トリガー（URL はハードコード）を追加すると、定期的なゲートとして実行できる。
+- 結果 JSON ファイル（`--screenshot` 指定時はフォーカスインジケーターのスクリーンショットも含む）はビルドアーティファクトとしてアップロードされる。
+- 繰り返し実行を高速化するには、インストール済みの `@playwright/test` バージョンをキーに `~/.cache/ms-playwright` をキャッシュする — パターンはこのリポジトリの `ci.yml` を参照。
+
 ## 使い方 — 関数 API（推奨）
 
 ページの遷移は呼び出し側で行い、関数は検査の実行・結果 JSON の書き出し・結果オブジェクトの

package/README.md

@@ -116,6 +116,63 @@ resolvable. If they are not found, the CLI exits 2 with an install hint.
 check is skipped with a `SKIPPED` message and does not affect the exit code.
 The other nine checks continue normally.
 
+### GitHub Actions
+
+Copy-paste workflow for a manually triggered audit. Violations fail the job
+(exit code 1); add `continue-on-error: true` to the audit step if you want the
+result to be informational only.
+
+```yaml
+# .github/workflows/a11y-audit.yml
+name: A11y Audit
+
+on:
+  workflow_dispatch:
+    inputs:
+      target_url:
+        description: 'Audit target URL'
+        required: true
+        default: 'https://example.com'
+
+jobs:
+  a11y-audit:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Setup Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: '24'
+
+      - name: Install audit package
+        run: npm install --no-save @a11y-skills/audit @playwright/test @axe-core/playwright
+
+      - name: Install Playwright browsers
+        run: npx playwright install --with-deps chromium
+
+      - name: Run a11y audit
+        run: npx a11y-audit --url "$TARGET_URL" --output-dir a11y-audit-results
+        env:
+          TARGET_URL: ${{ inputs.target_url }}
+
+      - name: Upload audit results
+        uses: actions/upload-artifact@v7
+        if: always()
+        with:
+          name: a11y-audit-results
+          path: a11y-audit-results/
+          retention-days: 30
+```
+
+Notes:
+
+- Add `push` / `schedule` triggers (with a hardcoded URL) to run the audit as
+  a recurring gate.
+- Result JSON files (and the focus-indicator screenshot when `--screenshot`
+  is set) are uploaded as a build artifact.
+- To speed up repeat runs, cache `~/.cache/ms-playwright` keyed on the
+  installed `@playwright/test` version — see this repository's `ci.yml` for
+  the pattern.
+
 ## Usage — function API (recommended)
 
 You navigate the page; the function runs the check, writes a result JSON, and

package/dist/cli.js

@@ -69,10 +69,12 @@ const CHECK_REGISTRY = [
     {
         name: 'orientation-check',
         kind: 'page',
-        async run({ page, outputDir }) {
+        ownsNavigation: true,
+        async run({ page, outputDir, url }) {
             const { runOrientationCheck } = await import('./playwright/runOrientationCheck.js');
             return (await runOrientationCheck({
                 page,
+                targetUrl: url,
                 outputDir,
             }));
         },
@@ -91,10 +93,12 @@ const CHECK_REGISTRY = [
     {
         name: 'time-limit-detector',
         kind: 'page',
-        async run({ page, outputDir }) {
+        ownsNavigation: true,
+        async run({ page, outputDir, url }) {
             const { runTimeLimitDetector } = await import('./playwright/runTimeLimitDetector.js');
             return (await runTimeLimitDetector({
                 page,
+                targetUrl: url,
                 outputDir,
             }));
         },
@@ -285,22 +289,24 @@ async function main() {
             if (check.kind === 'page') {
                 context = await browser.newContext();
                 page = await context.newPage();
-                // Use 'load' for file: URLs to avoid networkidle hanging
-                const waitUntil = url.startsWith('file:') ? 'load' : 'networkidle';
-                try {
-                    await page.goto(url, { waitUntil });
-                }
-                catch (navErr) {
-                    const msg = navErr instanceof Error ? navErr.message : String(navErr);
-                    process.stderr.write(`[${check.name}] Navigation failed: ${msg}\n`);
-                    summaries.push({
-                        name: check.name,
-                        status: 'ERROR',
-                        durationMs: Date.now() - start,
-                        error: msg,
-                    });
-                    hasErrors = true;
-                    continue;
+                if (!check.ownsNavigation) {
+                    // Use 'load' for file: URLs to avoid networkidle hanging
+                    const waitUntil = url.startsWith('file:') ? 'load' : 'networkidle';
+                    try {
+                        await page.goto(url, { waitUntil });
+                    }
+                    catch (navErr) {
+                        const msg = navErr instanceof Error ? navErr.message : String(navErr);
+                        process.stderr.write(`[${check.name}] Navigation failed: ${msg}\n`);
+                        summaries.push({
+                            name: check.name,
+                            status: 'ERROR',
+                            durationMs: Date.now() - start,
+                            error: msg,
+                        });
+                        hasErrors = true;
+                        continue;
+                    }
                 }
             }
             const result = await check.run({

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@a11y-skills/audit",
-  "version": "0.4.0",
+  "version": "0.4.1",
   "description": "Playwright + axe-core based WCAG 2.2 accessibility audit functions (axe, focus indicator, reflow, target size, text spacing, zoom, orientation, autocomplete, time limit, auto-play).",
   "type": "module",
   "license": "MIT",