From f8cf0cb7727a404a48144036726d766e17bfcbd4 Mon Sep 17 00:00:00 2001
From: Sergio Sisternes <sergio.sisternes@epam.com>
Date: Thu, 9 Apr 2026 19:20:44 +0200
Subject: [PATCH] fix(run): add explicit UTF-8 encoding to prompt file
 operations (#604)

All three open() calls in PromptCompiler and ScriptRunner that handle
.prompt.md files now specify encoding="utf-8", preventing
UnicodeDecodeError on systems where the default locale is not UTF-8
(e.g., Windows CP950).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 CHANGELOG.md                      |  1 +
 src/apm_cli/core/script_runner.py |  6 ++--
 tests/unit/test_script_runner.py  | 55 +++++++++++++++++++++++++++++++
 3 files changed, 59 insertions(+), 3 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 64f1f058..7db4be7e 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- Add explicit UTF-8 encoding to prompt file read/write operations to prevent `UnicodeDecodeError` on non-UTF-8 default locales (e.g., Windows CP950) (#604)
 - Propagate headers and environment variables through OpenCode MCP adapter with defensive copies to prevent mutation (#622)
 ### Changed
 
diff --git a/src/apm_cli/core/script_runner.py b/src/apm_cli/core/script_runner.py
index 87798c5c..fcb919fd 100644
--- a/src/apm_cli/core/script_runner.py
+++ b/src/apm_cli/core/script_runner.py
@@ -256,7 +256,7 @@ def _auto_compile_prompts(
             compiled_prompt_files.append(prompt_file)
 
             # Read the compiled content
-            with open(compiled_path, "r") as f:
+            with open(compiled_path, "r", encoding="utf-8") as f:
                 compiled_content = f.read().strip()
 
             # Check if this is a runtime command (copilot, codex, llm) before transformation
@@ -916,7 +916,7 @@ def compile(self, prompt_file: str, params: Dict[str, str]) -> str:
         # Now ensure compiled directory exists
         self.compiled_dir.mkdir(parents=True, exist_ok=True)
 
-        with open(prompt_path, "r") as f:
+        with open(prompt_path, "r", encoding="utf-8") as f:
             content = f.read()
 
         # Parse frontmatter and content
@@ -939,7 +939,7 @@ def compile(self, prompt_file: str, params: Dict[str, str]) -> str:
         output_path = self.compiled_dir / output_name
 
         # Write compiled content
-        with open(output_path, "w") as f:
+        with open(output_path, "w", encoding="utf-8") as f:
             f.write(compiled_content)
 
         return str(output_path)
diff --git a/tests/unit/test_script_runner.py b/tests/unit/test_script_runner.py
index 31c79b0a..af44baa8 100644
--- a/tests/unit/test_script_runner.py
+++ b/tests/unit/test_script_runner.py
@@ -325,6 +325,61 @@ def test_compile_file_not_found(self, mock_exists):
         with pytest.raises(FileNotFoundError, match="Prompt file 'nonexistent.prompt.md' not found"):
             self.compiler.compile("nonexistent.prompt.md", {})
 
+    def test_compile_utf8_content_with_cjk_characters(self):
+        """Test that prompt files with non-ASCII characters compile correctly.
+
+        Regression test for #604: UnicodeDecodeError on Windows CP950
+        when .prompt.md contains CJK or other non-ASCII characters.
+        """
+        tmp_dir = tempfile.mkdtemp()
+        try:
+            prompt_dir = Path(tmp_dir)
+            prompt_path = prompt_dir / "i18n.prompt.md"
+            cjk_content = (
+                "---\n"
+                "description: 国際化テスト\n"
+                "---\n"
+                "\n"
+                "你好世界！こんにちは ${input:name}！\n"
+                "Ünïcödé résumé naïve café"
+            )
+            prompt_path.write_text(cjk_content, encoding="utf-8")
+
+            compiler = PromptCompiler()
+            compiler.compiled_dir = prompt_dir / ".compiled"
+
+            result_path = compiler.compile(
+                str(prompt_path), {"name": "ユーザー"}
+            )
+
+            compiled = Path(result_path).read_text(encoding="utf-8")
+            assert "你好世界！こんにちは ユーザー！" in compiled
+            assert "Ünïcödé résumé naïve café" in compiled
+            # Frontmatter must be stripped
+            assert "---" not in compiled
+        finally:
+            shutil.rmtree(tmp_dir)
+
+    def test_compile_utf8_content_without_frontmatter(self):
+        """Test non-ASCII prompt without frontmatter compiles correctly."""
+        tmp_dir = tempfile.mkdtemp()
+        try:
+            prompt_dir = Path(tmp_dir)
+            prompt_path = prompt_dir / "simple_cjk.prompt.md"
+            prompt_path.write_text(
+                "Привет ${input:who}! 🚀", encoding="utf-8"
+            )
+
+            compiler = PromptCompiler()
+            compiler.compiled_dir = prompt_dir / ".compiled"
+
+            result_path = compiler.compile(str(prompt_path), {"who": "мир"})
+
+            compiled = Path(result_path).read_text(encoding="utf-8")
+            assert compiled == "Привет мир! 🚀"
+        finally:
+            shutil.rmtree(tmp_dir)
+
 
 class TestPromptCompilerDependencyDiscovery:
     """Test PromptCompiler dependency discovery functionality."""