chore(ci): fail CI when unapproved non-English text appears in source files

.github/workflows/ci.yml

@@ -32,6 +32,9 @@ jobs:
       - name: Verify action pins
         run: bash scripts/verify-action-pins.sh
 
+      - name: Verify sources contain no unapproved non-English text
+        run: go run scripts/verify-english-only.go
+
       - name: Check formatting
         run: |
           unformatted="$(gofmt -s -l .)"

AGENTS.md

@@ -22,8 +22,10 @@ open-code-review (`ocr`) is an AI-powered code review CLI tool written in Go (mo
 
 ## Code Style
 
-- After writing code, run `make check` to format and check the code.
-- `make check` runs: license check, `go mod tidy`, `gofmt -s -w .`, and `go vet`.
+- After writing code, run `make check`. It formats and tidies in place, so there is no need to run `gofmt` or `go vet` separately.
+- **Source files are written in English** — comments, identifiers and strings alike. `make english-check` enforces this in CI. It flags any letter outside ASCII, whichever the writing system (Han, kana, Hangul, Cyrillic, and equally the diacritics that spell German or Vietnamese), plus combining accents and fullwidth punctuation (`：`, `（`), which is easy to leave behind in an otherwise English sentence. Symbols and emoji (`─ → ≥ ✅`) pass, since they are not letters. Prose spelled entirely in ASCII (`Loeschen der Datei`, or a romanised transcription) takes a dictionary to spot and stays a matter for review.
+- **Translated prose has its own homes, none of them scanned.** `README.<locale>.md` and `CONTRIBUTING.<locale>.md` (`zh-CN`, `ja-JP`, `ko-KR`, `ru-RU`); the doc pages under `pages/src/content/docs/<locale>/` (`en`, `zh`, `ja`, `ru`, Markdown throughout); and the UI copy tables in `pages/src/i18n/<locale>.ts`. Markdown is out of scope by extension, so translations go there freely. The i18n tables are `.ts` and would be scanned, so they are exempt by prefix instead — translated UI strings belong in those tables rather than inline in a component.
+- **Two escape hatches for the exceptional case, narrower one preferred.** Append an `allow-non-english: <reason>` marker comment to the offending line — the right choice for a handful of lines, such as an encoding fixture or a language-switcher label, and it leaves the rest of the file protected. Only for a whole tree that is inherently non-English, add a prefix to `allowedPrefixes` in `scripts/verify-english-only.go`; it currently holds just `pages/src/i18n/` and `extensions/vscode/`, the latter temporary until the extension's Chinese comments are translated.
 
 ## Testing
 

Makefile

@@ -2,7 +2,7 @@
 	build-all dist sha256sum version-info \
 	build-linux-amd64 build-linux-arm64 build-darwin-amd64 build-darwin-arm64 \
 	build-windows-amd64 build-windows-arm64 \
-	license-check license-add
+	license-check license-add english-check
 
 BINARY_NAME := opencodereview
 GO          := go
@@ -64,7 +64,7 @@ fmt:
 vet:
 	LC_ALL=C $(GO) vet $(PACKAGES)
 
-check: license-check
+check: license-check english-check
 	$(GO) mod tidy
 	gofmt -s -w .
 	LC_ALL=C $(GO) vet $(PACKAGES)
@@ -73,6 +73,9 @@ check: license-check
 license-check:
 	@bash scripts/verify-license.sh
 
+english-check:
+	@$(GO) run scripts/verify-english-only.go
+
 license-add:
 	@bash scripts/add-license.sh
 

action.yml

@@ -36,7 +36,7 @@ inputs:
   language:
     description: >-
       Review output language, written via `ocr config set language`
-      (e.g. English, 中文). No env var exists for this.
+      (e.g. English, Chinese). No env var exists for this.
     required: false
     default: 'English'
   llm_timeout:

cmd/opencodereview/output_test.go

@@ -94,7 +94,7 @@ func TestSanitizeTerminal(t *testing.T) {
 		{"strips carriage return", "fake\rreal", "fakereal"},
 		{"empty string", "", ""},
 		{"only control chars", "\x1b\x07\x00\x7f", ""},
-		{"unicode preserved", "代码审查 レビュー 🔍", "代码审查 レビュー 🔍"},
+		{"unicode preserved", "代码审查 レビュー 🔍", "代码审查 レビュー 🔍"}, // allow-non-english: fixture asserts non-ASCII output is preserved verbatim
 		{"mixed safe and unsafe", "path\x1b[0m/file.go", "path[0m/file.go"},
 		{"strips C1 CSI (U+009B)", "before\u009bafter", "beforeafter"},
 		{"strips C1 OSC (U+009D)", "before\u009dafter", "beforeafter"},

cmd/opencodereview/session_cmd_test.go

@@ -287,11 +287,11 @@ func TestRunSessionShow_MissingID(t *testing.T) {
 }
 
 func TestTruncateUnicode(t *testing.T) {
-	got := truncate("错误原因：超过限制", 6)
+	got := truncate("错误原因：超过限制", 6) // allow-non-english: fixture exercises rune-boundary truncation
 	if !strings.HasSuffix(got, "…") {
 		t.Fatalf("expected ellipsis suffix, got %q", got)
 	}
-	if !strings.Contains(got, "错误") {
+	if !strings.Contains(got, "错误") { // allow-non-english: fixture exercises rune-boundary truncation
 		t.Fatalf("expected valid truncated unicode text, got %q", got)
 	}
 }

examples/gerrit_ci/post_review_test.py

@@ -134,10 +134,10 @@ def test_suggestion_without_existing(self):
         self.assertNotIn("**Suggestion:**", msg)
 
     def test_unicode_comment_roundtrip(self):
-        content = "空指针解引用：y 可能为 nil"
-        ri = build([comment(path="pkg/服务.go", content=content)])
-        self.assertIn("pkg/服务.go", ri["comments"])
-        self.assertIn(content, entry_of(ri, "pkg/服务.go")["message"])
+        content = "空指针解引用：y 可能为 nil"  # allow-non-english: fixture exercises UTF-8 comment bodies
+        ri = build([comment(path="pkg/服务.go", content=content)])  # allow-non-english: fixture exercises UTF-8 file paths
+        self.assertIn("pkg/服务.go", ri["comments"])  # allow-non-english: fixture exercises UTF-8 file paths
+        self.assertIn(content, entry_of(ri, "pkg/服务.go")["message"])  # allow-non-english: fixture exercises UTF-8 file paths
         self.assertEqual(json.loads(json.dumps(ri, ensure_ascii=False)), ri)
 
     def test_path_with_spaces(self):
@@ -539,15 +539,15 @@ def fake_urlopen(req, timeout=None):
     def test_preemptive_basic_auth_and_utf8_body(self):
         import base64
 
-        req, _parsed = self.post({"message": "空指针解引用：y 可能为 nil"})
+        req, _parsed = self.post({"message": "空指针解引用：y 可能为 nil"})  # allow-non-english: fixture exercises UTF-8 request payloads
         auth = req.get_header("Authorization")
         self.assertIsNotNone(auth, "Authorization header must be set preemptively")
         self.assertTrue(auth.startswith("Basic "))
         self.assertEqual(
             base64.b64decode(auth[len("Basic "):]).decode("utf-8"),
             "review-bot:s3cret-pass",
         )
-        self.assertIn("空指针解引用".encode("utf-8"), req.data)
+        self.assertIn("空指针解引用".encode("utf-8"), req.data)  # allow-non-english: fixture exercises UTF-8 request payloads
         self.assertIn("application/json", req.get_header("Content-type"))
 
     def test_xssi_response_parses(self):

internal/config/toolsconfig/tools.json

@@ -98,7 +98,7 @@
     "main_task": true,
     "definition": {
       "name": "file_read",
-      "description": "Use this tool to read file content when you need to get context for git diff. You can specify start_line and end_line to view specific parts of the file.\n\n**Line Range Strategy:**\n- Git diff hunk header provides guidance on how to get more relevant context.\n- Git diff hunk header \"@@-x,y +m,n@@\" indicates that the old file has y lines starting from line x, and the new file has n lines starting from line m.\n- For example, when you need to read 50 lines above and below the current changed code block in the new file, set start_line = m - 50, end_line = m + n + 50.\n\n**Example output:**\nFile：path/to/example.go (Total lines: 50)\nIS_TRUNCATED: false\nLINE_RANGE: 10-12\n// The following is the original content of the file\nfunc main() {\n  fmt.Println(\"Hello, World!\")\n}\n\n**Limitations:**\n- If the specified range exceeds 500 lines, only 500 lines will be returned with a truncation notice.\n- This tool can only read file content from the modified version (after changes) in git diff.",
+      "description": "Use this tool to read file content when you need to get context for git diff. You can specify start_line and end_line to view specific parts of the file.\n\n**Line Range Strategy:**\n- Git diff hunk header provides guidance on how to get more relevant context.\n- Git diff hunk header \"@@-x,y +m,n@@\" indicates that the old file has y lines starting from line x, and the new file has n lines starting from line m.\n- For example, when you need to read 50 lines above and below the current changed code block in the new file, set start_line = m - 50, end_line = m + n + 50.\n\n**Example output:**\nFile: path/to/example.go (Total lines: 50)\nIS_TRUNCATED: false\nLINE_RANGE: 10-12\n// The following is the original content of the file\nfunc main() {\n  fmt.Println(\"Hello, World!\")\n}\n\n**Limitations:**\n- If the specified range exceeds 500 lines, only 500 lines will be returned with a truncation notice.\n- This tool can only read file content from the modified version (after changes) in git diff.",
       "parameters": {
         "type": "object",
         "properties": {

internal/diff/git_test.go

@@ -81,7 +81,7 @@ func initRepoWithNonASCIIChange(t *testing.T) (string, string) {
 	runGitTest(t, repo, "config", "commit.gpgsign", "false")
 	runGitTest(t, repo, "config", "core.quotepath", "true")
 
-	relPath := "src/café/(authenticated)/文件.ts"
+	relPath := "src/café/(authenticated)/文件.ts" // allow-non-english: fixture exercises non-ASCII paths
 	file := filepath.Join(repo, filepath.FromSlash(relPath))
 	if err := os.MkdirAll(filepath.Dir(file), 0o755); err != nil {
 		t.Fatalf("create non-ASCII path: %v", err)
@@ -153,7 +153,7 @@ func TestWorkspaceDiffPreservesNonASCIIUntrackedPath(t *testing.T) {
 	repo, trackedPath := initRepoWithNonASCIIChange(t)
 	runGitTest(t, repo, "checkout", "--", trackedPath)
 
-	untrackedPath := "src/café/(authenticated)/新增.ts"
+	untrackedPath := "src/café/(authenticated)/新增.ts" // allow-non-english: fixture exercises non-ASCII paths
 	if err := os.WriteFile(filepath.Join(repo, filepath.FromSlash(untrackedPath)), []byte("untracked\n"), 0o644); err != nil {
 		t.Fatalf("write non-ASCII untracked file: %v", err)
 	}

internal/session/manifest_test.go

@@ -692,7 +692,7 @@ func TestSanitizeReasonTruncatesAndSingleLine(t *testing.T) {
 		t.Fatal("newlines not collapsed")
 	}
 	// Multibyte input must not be cut mid-rune.
-	multibyte := strings.Repeat("世", maxReasonLen+50)
+	multibyte := strings.Repeat("世", maxReasonLen+50) // allow-non-english: fixture exercises multibyte truncation
 	if !utf8.ValidString(sanitizeReason(multibyte)) {
 		t.Fatal("truncation produced invalid UTF-8")
 	}

internal/viewer/server_test.go

@@ -20,8 +20,8 @@ func TestTruncateText(t *testing.T) {
 		{"truncated with ellipsis", 3, "hello", "hel…"},
 		{"empty string", 5, "", ""},
 		{"n=0 always truncates non-empty", 0, "hi", "…"},
-		{"unicode shorter than n bytes", 20, "你好世界", "你好世界"},
-		{"unicode truncated at byte boundary", 6, "你好世界", "你好…"},
+		{"unicode shorter than n bytes", 20, "你好世界", "你好世界"},     // allow-non-english: fixture exercises rune-boundary truncation
+		{"unicode truncated at byte boundary", 6, "你好世界", "你好…"}, // allow-non-english: fixture exercises rune-boundary truncation
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

pages/src/components/Footer.tsx

@@ -10,9 +10,9 @@ import type { Language } from '../i18n/types';
 
 const LANG_OPTIONS: { value: Language; label: string }[] = [
   { value: 'en', label: 'English' },
-  { value: 'zh', label: '中文' },
-  { value: 'ja', label: '日本語' },
-  { value: 'ru', label: 'Русский' },
+  { value: 'zh', label: '中文' }, // allow-non-english: language options are labelled in their own language
+  { value: 'ja', label: '日本語' }, // allow-non-english: language options are labelled in their own language
+  { value: 'ru', label: 'Русский' }, // allow-non-english: language options are labelled in their own language
 ];
 
 const Footer: React.FC = () => {

pages/src/components/HeroSection.tsx

@@ -119,7 +119,7 @@ const terminalLines = [
   { num: 9, content: <span>&nbsp;</span> },
   { num: 10, content: <span style={{ color: TC.dim }}>─── internal/auth/login.go:42-45 ───</span> },
   { num: 11, content: <span style={{ color: TC.text }}>Consider using bcrypt cost factor ≥ 12 for password hashing.</span> },
-  { num: 12, content: <span className="terminal-cursor" style={{ color: TC.text }}>｜</span> },
+  { num: 12, content: <span className="terminal-cursor" style={{ color: TC.text }}>｜</span> }, // allow-non-english: fullwidth bar renders the terminal cursor
 ];
 
 const INSTALL_CHANNELS = [

pages/src/components/MarkdownRenderer.test.tsx

@@ -6,17 +6,24 @@ import { describe, expect, it } from 'vitest';
 import { LanguageProvider } from '../i18n';
 import MarkdownRenderer from './MarkdownRenderer';
 
+// The headings are Russian on purpose: explicit heading IDs exist for text that
+// cannot produce a usable ASCII slug on its own. Held in constants rather than
+// inline, because a marker comment on the JSX line would render as heading text.
+const H2 = 'Что делает навык'; // allow-non-english: fixture heading that cannot produce an ASCII slug
+const H4 = 'Публикация'; // allow-non-english: fixture heading that cannot produce an ASCII slug
+const CONTENT = `## ${H2} {#what-the-skill-does}\n\n#### ${H4} {#service-account}`;
+
 describe('MarkdownRenderer heading IDs', () => {
   it('renders an explicit heading ID without displaying its marker', () => {
     render(
       <LanguageProvider>
-        <MarkdownRenderer content={'## Что делает навык {#what-the-skill-does}\n\n#### Публикация {#service-account}'} />
+        <MarkdownRenderer content={CONTENT} />
       </LanguageProvider>,
     );
 
-    const heading = screen.getByRole('heading', { name: 'Что делает навык' });
+    const heading = screen.getByRole('heading', { name: H2 });
     expect(heading.getAttribute('id')).toBe('what-the-skill-does');
     expect(heading.textContent).not.toContain('{#what-the-skill-does}');
-    expect(screen.getByRole('heading', { name: 'Публикация', level: 4 }).getAttribute('id')).toBe('service-account');
+    expect(screen.getByRole('heading', { name: H4, level: 4 }).getAttribute('id')).toBe('service-account');
   });
 });

pages/src/components/Navbar.tsx

@@ -12,15 +12,15 @@ import type { Language } from '../i18n/types';
 
 const LANG_OPTIONS: { value: Language; label: string }[] = [
   { value: 'en', label: 'English' },
-  { value: 'zh', label: '中文' },
-  { value: 'ja', label: '日本語' },
-  { value: 'ru', label: 'Русский' },
+  { value: 'zh', label: '中文' }, // allow-non-english: language options are labelled in their own language
+  { value: 'ja', label: '日本語' }, // allow-non-english: language options are labelled in their own language
+  { value: 'ru', label: 'Русский' }, // allow-non-english: language options are labelled in their own language
 ];
 
 const LANG_BADGE: Record<Language, string> = {
   en: 'En',
-  zh: '中',
-  ja: 'あ',
+  zh: '中', // allow-non-english: single-glyph locale badge
+  ja: 'あ', // allow-non-english: single-glyph locale badge
   ru: 'Ru',
 };
 

pages/src/utils/headingId.test.ts

@@ -5,10 +5,14 @@ import { describe, expect, it } from 'vitest';
 import { extractHeadings } from './extractHeadings';
 import { parseExplicitHeadingId } from './headingId';
 
+// Russian on purpose: explicit heading IDs exist for text that cannot produce a
+// usable ASCII slug on its own.
+const HEADING = 'Что делает навк'; // allow-non-english: fixture heading that cannot produce an ASCII slug
+
 describe('explicit heading IDs', () => {
   it('separates a trailing explicit ID from the visible heading text', () => {
-    expect(parseExplicitHeadingId('Что делает навк {#what-the-skill-does}')).toEqual({
-      text: 'Что делает навк',
+    expect(parseExplicitHeadingId(`${HEADING} {#what-the-skill-does}`)).toEqual({
+      text: HEADING,
       id: 'what-the-skill-does',
     });
   });
@@ -18,8 +22,8 @@ describe('explicit heading IDs', () => {
   });
 
   it('uses the explicit ID in the table of contents without exposing its marker', () => {
-    expect(extractHeadings('## Что делает навк {#what-the-skill-does}')).toEqual([
-      { id: 'what-the-skill-does', text: 'Что делает навк', level: 2 },
+    expect(extractHeadings(`## ${HEADING} {#what-the-skill-does}`)).toEqual([
+      { id: 'what-the-skill-does', text: HEADING, level: 2 },
     ]);
   });
 });

scripts/github-actions/check-translation-sync.test.js

@@ -91,8 +91,8 @@ function testIdenticalStructurePasses() {
   // Same outline, DIFFERENT heading text (simulating translations). Must pass:
   // the check compares structure, not text.
   const en = readme(["## What is it?", "### Details", "## Usage"]);
-  const zh = readme(["## 这是什么？", "### 细节", "## 使用方法"]);
-  const ja = readme(["## これは何ですか？", "### 詳細", "## 使い方"]);
+  const zh = readme(["## 这是什么？", "### 细节", "## 使用方法"]); // allow-non-english: fixture mimics translated README headings
+  const ja = readme(["## これは何ですか？", "### 詳細", "## 使い方"]); // allow-non-english: fixture mimics translated README headings
   const { ok, errors } = compareReadmeStructures([
     { name: "README.md", content: en },
     { name: "README.zh-CN.md", content: zh },