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 are English-only
+        run: go run scripts/verify-cjk.go
+
       - name: Check formatting
         run: |
           unformatted="$(gofmt -s -l .)"

AGENTS.md

@@ -23,7 +23,9 @@ 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`.
+- `make check` runs: license check, CJK check, `go mod tidy`, `gofmt -s -w .`, and `go vet`.
+- Source files are English-only — comments, identifiers and strings alike. Translated prose belongs in `README.<locale>.md`, `pages/src/content/docs/<locale>/` or an i18n table. `make cjk-check` enforces this in CI; it also rejects fullwidth punctuation (`：`, `（`), which is easy to leave behind in an otherwise English sentence.
+- When CJK text is intentional (an encoding fixture, a language-switcher label), append an `allow-cjk: <reason>` marker comment to that line rather than widening the allowlist in `scripts/verify-cjk.go`.
 
 ## 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 cjk-check
 
 BINARY_NAME := opencodereview
 GO          := go
@@ -64,7 +64,7 @@ fmt:
 vet:
 	LC_ALL=C $(GO) vet $(PACKAGES)
 
-check: license-check
+check: license-check cjk-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
 
+cjk-check:
+	@$(GO) run scripts/verify-cjk.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-cjk: 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-cjk: 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-cjk: 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-cjk: fixture exercises UTF-8 comment bodies
+        ri = build([comment(path="pkg/服务.go", content=content)])  # allow-cjk: fixture exercises UTF-8 file paths
+        self.assertIn("pkg/服务.go", ri["comments"])  # allow-cjk: fixture exercises UTF-8 file paths
+        self.assertIn(content, entry_of(ri, "pkg/服务.go")["message"])  # allow-cjk: 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-cjk: 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-cjk: 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-cjk: 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-cjk: 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-cjk: 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-cjk: fixture exercises rune-boundary truncation
+		{"unicode truncated at byte boundary", 6, "你好世界", "你好…"}, // allow-cjk: fixture exercises rune-boundary truncation
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

pages/src/components/Footer.tsx

@@ -10,8 +10,8 @@ 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: 'zh', label: '中文' }, // allow-cjk: language options are labelled in their own language
+  { value: 'ja', label: '日本語' }, // allow-cjk: language options are labelled in their own language
   { value: 'ru', label: 'Русский' },
 ];
 

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-cjk: fullwidth bar renders the terminal cursor
 ];
 
 const INSTALL_CHANNELS = [

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: 'zh', label: '中文' }, // allow-cjk: language options are labelled in their own language
+  { value: 'ja', label: '日本語' }, // allow-cjk: language options are labelled in their own language
   { value: 'ru', label: 'Русский' },
 ];
 
 const LANG_BADGE: Record<Language, string> = {
   en: 'En',
-  zh: '中',
-  ja: 'あ',
+  zh: '中', // allow-cjk: single-glyph locale badge
+  ja: 'あ', // allow-cjk: single-glyph locale badge
   ru: 'Ru',
 };
 

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-cjk: fixture mimics translated README headings
+  const ja = readme(["## これは何ですか？", "### 詳細", "## 使い方"]); // allow-cjk: fixture mimics translated README headings
   const { ok, errors } = compareReadmeStructures([
     { name: "README.md", content: en },
     { name: "README.zh-CN.md", content: zh },