From ef0f4f7ef49c6fc0a61a9a9a1d783ad0713554ee Mon Sep 17 00:00:00 2001 From: jxxghp Date: Tue, 1 Sep 2026 17:51:15 +0800 Subject: [PATCH] feat(agent): improve Windows shell compatibility --- app/agent/llm/capability.py | 24 +- app/agent/mcp.py | 12 +- app/agent/prompt/System Core Prompt.txt | 1 + app/agent/prompt/__init__.py | 4 + app/agent/shell.py | 179 +++++++++++ app/agent/tools/impl/_terminal_session.py | 33 +- app/agent/tools/impl/execute_command.py | 25 +- app/agent/tools/impl/service.py | 3 + docs/rules/05-architecture.md | 2 + tests/test_agent_shell.py | 361 ++++++++++++++++++++++ 10 files changed, 610 insertions(+), 34 deletions(-) create mode 100644 app/agent/shell.py create mode 100644 tests/test_agent_shell.py diff --git a/app/agent/llm/capability.py b/app/agent/llm/capability.py index c417b37d9..61d63fc8e 100644 --- a/app/agent/llm/capability.py +++ b/app/agent/llm/capability.py @@ -12,11 +12,11 @@ from pathlib import Path from typing import Any, Dict, Optional from uuid import uuid4 -from app.application.notification import get_notification_configs -from app.runtime.settings import get_runtime_setting - -from app.runtime.log import logger from app.adapters.network.http import RequestUtils +from app.agent.shell import agent_text_subprocess_kwargs +from app.application.notification import get_notification_configs +from app.runtime.log import logger +from app.runtime.settings import get_runtime_setting class AgentCapabilityProvider(ABC): @@ -286,7 +286,13 @@ class OpenAIChatAudioProvider(AudioCapabilityProvider): "wav", str(output_path), ] - result = subprocess.run(cmd, capture_output=True, text=True, check=False) + result = subprocess.run( + cmd, + capture_output=True, + text=True, + check=False, + **agent_text_subprocess_kwargs(), + ) if result.returncode != 0 or not output_path.exists(): logger.warning( "%s STT 音频转 WAV 失败: returncode=%s, stderr=%s", @@ -354,7 +360,13 @@ class OpenAIChatAudioProvider(AudioCapabilityProvider): "libopus", str(output_path), ] - result = subprocess.run(cmd, capture_output=True, text=True, check=False) + result = subprocess.run( + cmd, + capture_output=True, + text=True, + check=False, + **agent_text_subprocess_kwargs(), + ) if result.returncode != 0 or not output_path.exists(): logger.warning( "%s TTS 音频转 Opus 失败，将使用 WAV 原文件: returncode=%s, stderr=%s", diff --git a/app/agent/mcp.py b/app/agent/mcp.py index d04f4861e..0386bf8e2 100644 --- a/app/agent/mcp.py +++ b/app/agent/mcp.py @@ -5,13 +5,14 @@ from __future__ import annotations import asyncio import hashlib import json -import os import re import uuid from dataclasses import dataclass from typing import Any, Optional from urllib.parse import urljoin +from app.adapters.network.http import AsyncRequestUtils +from app.agent.shell import build_agent_subprocess_env from app.application.configuration import get_configured_system_config from app.runtime.log import logger from app.schemas.agent import ( @@ -20,7 +21,6 @@ from app.schemas.agent import ( AgentMcpServerToolInfo, ) from app.schemas.types import SystemConfigKey -from app.adapters.network.http import AsyncRequestUtils MCP_PROTOCOL_VERSION = "2025-11-25" MCP_CLIENT_NAME = "MoviePilot Agent" @@ -202,8 +202,7 @@ class _StdioMcpSession: """启动 stdio MCP 子进程。""" if not self.server.command: raise RuntimeError("stdio MCP 服务器缺少启动命令") - env = os.environ.copy() - env.update(self.server.env or {}) + env = build_agent_subprocess_env(self.server.env) self.process = await asyncio.create_subprocess_exec( self.server.command, *(self.server.args or []), @@ -241,7 +240,10 @@ class _StdioMcpSession: line = await self.process.stderr.readline() if not line: break - logger.debug(f"MCP stdio[{self.server.name}] stderr: {line.decode(errors='replace').strip()}") + logger.debug( + f"MCP stdio[{self.server.name}] stderr: " + f"{line.decode('utf-8', errors='replace').strip()}" + ) except asyncio.CancelledError: return diff --git a/app/agent/prompt/System Core Prompt.txt b/app/agent/prompt/System Core Prompt.txt index 84f9b4075..3abd5b3a2 100644 --- a/app/agent/prompt/System Core Prompt.txt +++ b/app/agent/prompt/System Core Prompt.txt @@ -71,6 +71,7 @@ You act as a proactive agent. Your goal is to fully resolve the user's media-rel - Read the relevant file before changing it, then pick the editing tool by scope. Use `apply_patch` when one logical change spans multiple files, adds new files, or deletes files: submit a single patch wrapped in `*** Begin Patch` / `*** End Patch` with `*** Add File:`, `*** Update File:`, and `*** Delete File:` sections; every context and removed line must match the current content exactly, and the whole patch is validated before any file is written. Use `edit_file` for a single localized exact replacement within one already-read file; make `old_text` unique with enough surrounding context, and use `replace_all=true` only when every match must change. Use `write_file` for one standalone new file; set `overwrite=true` only for an intentional full rewrite, and use `read_file(include_metadata=true)` plus `expected_sha256` when preserving the previously read version matters. - When implementation depends on a Python or Node.js API, first identify the installed or locked dependency version from environment metadata, requirements, package manifests, lockfiles, local source, and type declarations. Use `rg` against the relevant package directory, `.venv`, or `node_modules` instead of scanning the entire project without bounds. If local evidence is insufficient, use `search_web` and then `browse_webpage` to read the matching version of the official documentation. Do not guess signatures from memory, mix examples from incompatible versions, or install a package only to inspect its API. - Use structured file tools for source edits because they enforce file access boundaries and conflict checks. Never use shell redirection, inline scripts, or another tool to bypass a file-tool permission denial. +- On Windows, follow the command environment injected in `` for the entire task. If Git Bash is selected, use Git Bash/POSIX syntax and prefer `git` for repository work. Otherwise, when PowerShell 7 is selected, use `pwsh` syntax consistently. Never mix cmd.exe, Windows PowerShell, PowerShell 7, and POSIX syntax in one workflow, and treat command and file text as UTF-8. - Use `execute_command` for administrator-only multi-file diagnostics, tests, Git, service operations, SSH, or an exact command the user requested. Use `action=run` for short bounded commands. Use `action=start` for long-running or interactive commands, including SSH; then continue with `read`, `wait`, `write`, or `kill` using the returned `session_id`. Do not start a background session for a short command that can finish within `action=run`. diff --git a/app/agent/prompt/__init__.py b/app/agent/prompt/__init__.py index 445ef5c27..ccc0126ae 100644 --- a/app/agent/prompt/__init__.py +++ b/app/agent/prompt/__init__.py @@ -11,6 +11,7 @@ import yaml from app.adapters.system.host import SystemUtils from app.agent.llm.capability import AgentCapabilityManager +from app.agent.shell import resolve_agent_shell from app.runtime.log import logger from app.runtime.settings import get_runtime_setting from app.schemas.notification import ( @@ -265,6 +266,9 @@ class PromptManager: f"- 运行环境: {SystemUtils.platform} {'docker' if SystemUtils.is_docker() else ''}", "- 详细运行状态和数据库通过 `query_doctor_report` 或 `execute_command` 查询；配置值先加载对应 Skill，再通过 `moviepilot_api` 的配置 operation 查询。", ] + shell = resolve_agent_shell() + if shell: + info_lines.append(shell.prompt_guidance()) path_lines = self._get_runtime_path_lines() if path_lines: info_lines.extend( diff --git a/app/agent/shell.py b/app/agent/shell.py new file mode 100644 index 000000000..3d7a7436c --- /dev/null +++ b/app/agent/shell.py @@ -0,0 +1,179 @@ +"""Agent 命令解释器选择与子进程编码策略。""" + +from __future__ import annotations + +import os +import shutil +from dataclasses import dataclass +from pathlib import Path +from typing import Any, Callable, Mapping, Optional + +WINDOWS_UTF8_ENVIRONMENT = { + "LANG": "C.UTF-8", + "LC_ALL": "C.UTF-8", + "PYTHONIOENCODING": "utf-8", + "PYTHONUTF8": "1", +} + +_POWERSHELL_UTF8_PREFIX = ( + "$utf8 = [System.Text.UTF8Encoding]::new($false); " + "[Console]::InputEncoding = $utf8; " + "[Console]::OutputEncoding = $utf8; " + "$OutputEncoding = $utf8; " + "$PSDefaultParameterValues['*:Encoding'] = 'utf8'; " + "chcp.com 65001 | Out-Null;" +) + +CommandFinder = Callable[[str], Optional[str]] + + +@dataclass(frozen=True) +class AgentShell: + """描述 Windows Agent 实际使用的命令解释器及模型指导。""" + + kind: str + executable: str + arguments: tuple[str, ...] + command_prefix: str = "" + git_available: bool = False + + def build_argv(self, command: str) -> list[str]: + """生成无需再次经过系统默认 shell 的命令参数。""" + command_text = f"{self.command_prefix}\n{command}" if self.command_prefix else command + return [self.executable, *self.arguments, command_text] + + def prompt_guidance(self) -> str: + """返回不包含可执行文件绝对路径的 Windows 命令指导。""" + if self.kind == "git-bash": + return ( + "- Windows Agent 命令环境: Git Bash（UTF-8）。所有 `execute_command` " + "命令统一使用 Git Bash/POSIX 语法；涉及仓库、源码状态或版本控制时优先使用 " + "`git`，不要混用 cmd.exe 或 PowerShell 语法。" + ) + if self.kind == "pwsh": + git_guidance = "；仓库和版本控制操作仍优先使用 `git`" if self.git_available else "" + return ( + "- Windows Agent 命令环境: PowerShell 7 (`pwsh`, UTF-8)。所有 " + f"`execute_command` 命令统一使用 PowerShell 语法{git_guidance}，不要混用 " + "cmd.exe、Windows PowerShell 或 POSIX shell 语法。" + ) + git_guidance = "；仓库和版本控制操作优先使用 `git`" if self.git_available else "" + return ( + "- Windows Agent 命令环境: cmd.exe UTF-8 兼容回退" + f"{git_guidance}。仅在 Git Bash 和 PowerShell 7 不可用时使用 cmd 语法。" + ) + + +def _find_git_bash(git_path: str, command_finder: CommandFinder) -> Optional[str]: + """从 Git for Windows 常见布局中定位同一安装目录下的 bash.exe。""" + git_executable = Path(git_path) + git_root = git_executable.parent.parent + candidates = ( + git_executable.parent / "bash.exe", + git_root / "bin" / "bash.exe", + git_root / "usr" / "bin" / "bash.exe", + ) + for candidate in candidates: + if candidate.is_file(): + return str(candidate) + + bash_path = command_finder("bash") or command_finder("bash.exe") + if not bash_path: + return None + bash_executable = Path(bash_path) + if bash_executable.parent == git_executable.parent: + return str(bash_executable) + if bash_executable.parent.parent == git_root: + return str(bash_executable) + return None + + +def resolve_agent_shell( + *, + platform_name: Optional[str] = None, + command_finder: Optional[CommandFinder] = None, + environment: Optional[Mapping[str, str]] = None, +) -> Optional[AgentShell]: + """ + 解析 Windows Agent 命令解释器，非 Windows 返回 None 以保留原有 shell 行为。 + + :param platform_name: 用于测试覆盖的 os.name，默认读取当前平台 + :param command_finder: 命令定位函数，默认使用 shutil.which + :param environment: 用于读取 COMSPEC 的环境变量 + :return: Windows 命令解释器策略，非 Windows 返回 None + """ + if (platform_name or os.name) != "nt": + return None + + finder = command_finder or shutil.which + current_environment = environment if environment is not None else os.environ + git_path = finder("git") or finder("git.exe") + if git_path: + bash_path = _find_git_bash(git_path, finder) + if bash_path: + return AgentShell( + kind="git-bash", + executable=bash_path, + arguments=("--noprofile", "--norc", "-lc"), + git_available=True, + ) + + pwsh_path = finder("pwsh") or finder("pwsh.exe") + if pwsh_path: + return AgentShell( + kind="pwsh", + executable=pwsh_path, + arguments=("-NoLogo", "-NoProfile", "-Command"), + command_prefix=_POWERSHELL_UTF8_PREFIX, + git_available=bool(git_path), + ) + + cmd_path = ( + current_environment.get("COMSPEC") + or finder("cmd") + or finder("cmd.exe") + or "cmd.exe" + ) + return AgentShell( + kind="cmd", + executable=cmd_path, + arguments=("/d", "/s", "/c"), + command_prefix="chcp 65001>nul &", + git_available=bool(git_path), + ) + + +def build_agent_subprocess_env( + overrides: Optional[Mapping[str, Any]] = None, + *, + platform_name: Optional[str] = None, +) -> dict[str, str]: + """ + 构造 Agent 子进程环境，并在 Windows 上强制覆盖为 UTF-8 编解码。 + + :param overrides: 调用方附加的环境变量 + :param platform_name: 用于测试覆盖的 os.name，默认读取当前平台 + :return: 可直接传给 subprocess 的字符串环境变量 + """ + environment = os.environ.copy() + for key, value in (overrides or {}).items(): + if value is not None: + environment[str(key)] = str(value) + if (platform_name or os.name) == "nt": + # Python 会把该变量的任意非空值视为启用，必须移除而不是写入 "0"。 + environment.pop("PYTHONLEGACYWINDOWSSTDIO", None) + environment.update(WINDOWS_UTF8_ENVIRONMENT) + return environment + + +def agent_text_subprocess_kwargs( + *, platform_name: Optional[str] = None +) -> dict[str, Any]: + """为 Windows Agent 文本子进程补充 UTF-8 解码参数，POSIX 返回空参数。""" + if (platform_name or os.name) != "nt": + return {} + return { + "encoding": "utf-8", + "errors": "replace", + "env": build_agent_subprocess_env(platform_name="nt"), + } diff --git a/app/agent/tools/impl/_terminal_session.py b/app/agent/tools/impl/_terminal_session.py index eb3caf9fe..425113c11 100644 --- a/app/agent/tools/impl/_terminal_session.py +++ b/app/agent/tools/impl/_terminal_session.py @@ -13,6 +13,7 @@ from dataclasses import dataclass, field from pathlib import Path from typing import Any, Optional +from app.agent.shell import build_agent_subprocess_env, resolve_agent_shell from app.agent.tools.impl._command_safety import validate_command_safety from app.runtime.log import logger from app.runtime.settings import get_runtime_setting @@ -163,15 +164,8 @@ class _TerminalSessionManager: @staticmethod def _build_env(env: Optional[dict[str, Any]]) -> dict[str, str]: - """合并环境变量，并把值稳定转换为字符串。""" - merged_env = os.environ.copy() - if not env: - return merged_env - for key, value in env.items(): - if value is None: - continue - merged_env[str(key)] = str(value) - return merged_env + """合并环境变量，Windows 强制覆盖为 UTF-8 编解码。""" + return build_agent_subprocess_env(env) @staticmethod def _validate_command(command: str, *, confirmed: bool = False) -> None: @@ -307,12 +301,21 @@ class _TerminalSessionManager: self, command: str, cwd: str, env: dict[str, str] ) -> _TerminalSession: """通过普通 stdin/stdout/stderr 管道启动命令会话。""" - process = await asyncio.create_subprocess_shell( - command, - cwd=cwd, - env=env, - **self._pipe_subprocess_kwargs(), - ) + shell = resolve_agent_shell() + if shell: + process = await asyncio.create_subprocess_exec( + *shell.build_argv(command), + cwd=cwd, + env=env, + **self._pipe_subprocess_kwargs(), + ) + else: + process = await asyncio.create_subprocess_shell( + command, + cwd=cwd, + env=env, + **self._pipe_subprocess_kwargs(), + ) session = _TerminalSession( session_id=f"term_{uuid.uuid4().hex[:12]}", command=command, diff --git a/app/agent/tools/impl/execute_command.py b/app/agent/tools/impl/execute_command.py index a3bc9568e..70848f74d 100644 --- a/app/agent/tools/impl/execute_command.py +++ b/app/agent/tools/impl/execute_command.py @@ -14,18 +14,18 @@ from typing import Any, Literal, Optional, TextIO, Type from pydantic import BaseModel, Field -from app.agent.tools.impl._command_safety import validate_command_safety +from app.agent.shell import build_agent_subprocess_env, resolve_agent_shell from app.agent.tools.base import MoviePilotTool -from app.agent.tools.tags import ToolTag +from app.agent.tools.impl._command_safety import validate_command_safety from app.agent.tools.impl._terminal_session import ( TERMINAL_DEFAULT_READ_BYTES, TERMINAL_MAX_READ_BYTES, TERMINAL_WAIT_DEFAULT_MS, get_terminal_session_manager, ) +from app.agent.tools.tags import ToolTag from app.runtime.log import logger - DEFAULT_TIMEOUT_SECONDS = 60 MAX_TIMEOUT_SECONDS = 300 MAX_OUTPUT_PREVIEW_BYTES = 32 * 1024 @@ -466,11 +466,20 @@ class ExecuteCommandTool(MoviePilotTool): normalized_timeout, timeout_note = self._normalize_timeout(timeout) async with _command_semaphore: - process = await asyncio.create_subprocess_shell( - command, - cwd=cwd, - **self._subprocess_kwargs(), - ) + shell = resolve_agent_shell() + if shell: + process = await asyncio.create_subprocess_exec( + *shell.build_argv(command), + cwd=cwd, + env=build_agent_subprocess_env(), + **self._subprocess_kwargs(), + ) + else: + process = await asyncio.create_subprocess_shell( + command, + cwd=cwd, + **self._subprocess_kwargs(), + ) output = _CommandOutput(preview_limit_bytes=MAX_OUTPUT_PREVIEW_BYTES) wait_task = asyncio.create_task(process.wait()) reader_tasks = [ diff --git a/app/agent/tools/impl/service.py b/app/agent/tools/impl/service.py index 747f70d0a..f115ad815 100644 --- a/app/agent/tools/impl/service.py +++ b/app/agent/tools/impl/service.py @@ -12,6 +12,7 @@ from typing import Any, ClassVar, Dict, Optional, Type from pydantic import BaseModel, ConfigDict, Field +from app.agent.shell import agent_text_subprocess_kwargs from app.agent.tools.base import MoviePilotTool from app.agent.tools.tags import ToolTag from app.runtime.settings import get_runtime_setting @@ -325,6 +326,7 @@ def _run_service_script( capture_output=True, text=True, check=False, + **agent_text_subprocess_kwargs(), ) return _parse_script_payload(completed.stdout) @@ -505,6 +507,7 @@ def _run_database_script(arguments: Dict[str, Any], *, root_path: Path) -> dict[ capture_output=True, text=True, check=False, + **agent_text_subprocess_kwargs(), ) if completed.returncode != 0: error_text = completed.stderr.strip() or completed.stdout.strip() or "数据库脚本执行失败" diff --git a/docs/rules/05-architecture.md b/docs/rules/05-architecture.md index d2bab91ce..d64c33765 100644 --- a/docs/rules/05-architecture.md +++ b/docs/rules/05-architecture.md @@ -86,6 +86,7 @@ to make the directory tree look symmetrical. | `app/agent/lifecycle.py` | Runtime generation admission, initialization, idle-session collection and bounded manager shutdown | | `app/agent/tasks.py` | Background prompts, durable scheduled-task execution and heartbeat wakeups | | `app/agent/orchestrator.py` | One `MoviePilotAgent` execution instance: prompt/tool/middleware assembly, model invocation, streaming and per-agent state | +| `app/agent/shell.py` | Agent command-shell selection and subprocess text-encoding policy; Windows prefers Git Bash, then PowerShell 7, while POSIX keeps native shell/PTY behavior | | `app/agent/policy/api.py` | Fixed `moviepilot_api` operation registry, HTTP route templates and per-operation authorization/effect policy; no arbitrary URL or method input | | `app/agent/policy/mcp.py` | Generated external MCP input-contract builder for the fixed API registry; owns exact English oneOf parameter projection, not runtime authorization | | `app/agent/tools/impl/service.py` | Admin-only external MCP wrappers for downloader, media-server and database Skill scripts; synchronous scripts run only through the Agent blocking executor | @@ -977,6 +978,7 @@ driven workflow registration. | `app/agent/lifecycle.py` | Agent manager admission, startup, idle collection and bounded shutdown owner | | `app/agent/tasks.py` | Background prompt, scheduled task and heartbeat execution owner | | `app/agent/orchestrator.py` | Per-session `MoviePilotAgent` execution and LLM/tool/middleware orchestration only | +| `app/agent/shell.py` | Agent command-shell selection and subprocess UTF-8 policy, with Windows-only Git Bash/PowerShell 7 routing | | `app/agent/loader.py` | Agent-specific capability discovery and canonical entrypoint/service materialization; reuses the generic Capability Runtime while keeping Agent ownership under `app/agent/` | | `app/agent/__init__.py` | Implementation-free package root; exact historical Agent symbols are supplied by the Compat overlay only, while host callers import `orchestrator.py` or the relevant owner directly | | `app/agent/llm/__init__.py` | Implementation-free package root; only the verified historical `LLMHelper` symbol is supplied by exact Compat routing | diff --git a/tests/test_agent_shell.py b/tests/test_agent_shell.py new file mode 100644 index 000000000..f6d1f460f --- /dev/null +++ b/tests/test_agent_shell.py @@ -0,0 +1,361 @@ +import asyncio +import json +import subprocess +from pathlib import Path +from types import SimpleNamespace +from unittest.mock import AsyncMock, patch + +import pytest + +from app.agent.llm.capability import OpenAIChatAudioProvider +from app.agent.mcp import _StdioMcpSession +from app.agent.prompt import PromptManager +from app.agent.shell import ( + AgentShell, + agent_text_subprocess_kwargs, + build_agent_subprocess_env, + resolve_agent_shell, +) +from app.agent.tools.impl._terminal_session import _TerminalSessionManager +from app.agent.tools.impl.execute_command import ExecuteCommandTool +from app.agent.tools.impl.service import _run_service_script +from app.schemas.agent import AgentMcpServerConfig + + +def _finder(paths: dict[str, str]): + """构造只返回测试声明命令路径的 which 替身。""" + return lambda command: paths.get(command) + + +def _fake_process(stdout: bytes = b"", stderr: bytes = b""): + """构造满足 Agent 异步子进程读取合同的最小进程替身。""" + stdout_reader = asyncio.StreamReader() + stdout_reader.feed_data(stdout) + stdout_reader.feed_eof() + stderr_reader = asyncio.StreamReader() + stderr_reader.feed_data(stderr) + stderr_reader.feed_eof() + + process = SimpleNamespace( + pid=12345, + returncode=0, + stdin=None, + stdout=stdout_reader, + stderr=stderr_reader, + terminate=lambda: None, + kill=lambda: None, + ) + + async def wait() -> int: + """立即返回成功退出码。""" + return 0 + + process.wait = wait + return process + + +def test_windows_shell_prefers_git_bash(tmp_path: Path) -> None: + """Windows 检测到 Git for Windows 时应优先使用同目录 Git Bash。""" + git_root = tmp_path / "Git" + git_path = git_root / "cmd" / "git.exe" + bash_path = git_root / "bin" / "bash.exe" + git_path.parent.mkdir(parents=True) + bash_path.parent.mkdir(parents=True) + git_path.touch() + bash_path.touch() + + shell = resolve_agent_shell( + platform_name="nt", + command_finder=_finder({"git": str(git_path), "pwsh": "ignored"}), + environment={}, + ) + + assert shell is not None + assert shell.kind == "git-bash" + assert shell.build_argv("git status") == [ + str(bash_path), + "--noprofile", + "--norc", + "-lc", + "git status", + ] + assert "Git Bash" in shell.prompt_guidance() + assert str(bash_path) not in shell.prompt_guidance() + + +def test_windows_shell_uses_pwsh_when_git_is_unavailable() -> None: + """Windows 没有 Git 时应统一选择 PowerShell 7，而不是旧 powershell.exe。""" + shell = resolve_agent_shell( + platform_name="nt", + command_finder=_finder({"pwsh": "C:/PowerShell/pwsh.exe"}), + environment={}, + ) + + assert shell is not None + assert shell.kind == "pwsh" + argv = shell.build_argv("Get-ChildItem") + assert argv[:4] == [ + "C:/PowerShell/pwsh.exe", + "-NoLogo", + "-NoProfile", + "-Command", + ] + assert "[Console]::OutputEncoding" in argv[-1] + assert "chcp.com 65001" in argv[-1] + assert argv[-1].endswith("Get-ChildItem") + assert "PowerShell 7" in shell.prompt_guidance() + + +def test_windows_shell_keeps_git_priority_when_only_pwsh_can_host_commands() -> None: + """Git 存在但无法定位 Git Bash 时，pwsh 仍应提示仓库操作优先使用 git。""" + shell = resolve_agent_shell( + platform_name="nt", + command_finder=_finder( + {"git": "C:/Portable/git.exe", "pwsh": "C:/PowerShell/pwsh.exe"} + ), + environment={}, + ) + + assert shell is not None + assert shell.kind == "pwsh" + assert shell.git_available is True + assert "优先使用 `git`" in shell.prompt_guidance() + + +def test_windows_shell_uses_utf8_cmd_fallback_without_git_or_pwsh() -> None: + """Git Bash 与 pwsh 都不可用时应保留显式 UTF-8 的 cmd 兼容回退。""" + shell = resolve_agent_shell( + platform_name="nt", + command_finder=_finder({}), + environment={"COMSPEC": "C:/Windows/System32/cmd.exe"}, + ) + + assert shell is not None + assert shell.kind == "cmd" + assert shell.build_argv("dir")[-1] == "chcp 65001>nul &\ndir" + + +def test_posix_shell_policy_preserves_native_shell_and_environment(monkeypatch) -> None: + """Linux/macOS 不应被 Windows shell 选择或 UTF-8 环境覆盖改变。""" + monkeypatch.setenv("PYTHONIOENCODING", "locale-default") + + assert resolve_agent_shell(platform_name="posix") is None + environment = build_agent_subprocess_env( + {"CUSTOM": "value"}, platform_name="posix" + ) + assert environment["PYTHONIOENCODING"] == "locale-default" + assert environment["CUSTOM"] == "value" + assert agent_text_subprocess_kwargs(platform_name="posix") == {} + + +def test_windows_subprocess_environment_forces_utf8_after_overrides() -> None: + """Windows 调用方不能通过自定义环境重新引入本地默认编码。""" + environment = build_agent_subprocess_env( + { + "PYTHONIOENCODING": "cp936", + "PYTHONLEGACYWINDOWSSTDIO": "1", + "PYTHONUTF8": "0", + "LANG": "zh_CN.GBK", + }, + platform_name="nt", + ) + + assert environment["PYTHONIOENCODING"] == "utf-8" + assert environment["PYTHONUTF8"] == "1" + assert environment["LANG"] == "C.UTF-8" + assert environment["LC_ALL"] == "C.UTF-8" + assert "PYTHONLEGACYWINDOWSSTDIO" not in environment + + +def test_prompt_injects_selected_windows_shell_without_executable_path() -> None: + """模型提示词应知道实际 Windows shell，但不能暴露安装绝对路径。""" + shell = AgentShell( + kind="pwsh", + executable="C:/secret/location/pwsh.exe", + arguments=("-Command",), + ) + with patch("app.agent.prompt.resolve_agent_shell", return_value=shell): + moviepilot_info = PromptManager()._get_moviepilot_info() + + assert "PowerShell 7" in moviepilot_info + assert "C:/secret/location/pwsh.exe" not in moviepilot_info + + +@pytest.mark.anyio +async def test_execute_command_uses_selected_windows_shell_and_utf8_environment() -> None: + """一次性命令在 Windows 策略下应绕过隐式 cmd 并传入 UTF-8 环境。""" + shell = AgentShell( + kind="pwsh", + executable="pwsh.exe", + arguments=("-Command",), + command_prefix="utf8-prefix;", + ) + process = _fake_process(stdout="中文输出".encode("utf-8")) + create_exec = AsyncMock(return_value=process) + tool = ExecuteCommandTool(session_id="session", user_id="user") + + with ( + patch( + "app.agent.tools.impl.execute_command.resolve_agent_shell", + return_value=shell, + ), + patch( + "app.agent.tools.impl.execute_command.build_agent_subprocess_env", + return_value={"PYTHONUTF8": "1"}, + ), + patch( + "app.agent.tools.impl.execute_command.asyncio.create_subprocess_exec", + create_exec, + ), + patch( + "app.agent.tools.impl.execute_command.asyncio.create_subprocess_shell" + ) as create_shell, + ): + result = await tool.run(action="run", command="Get-Location", timeout=1) + + assert "中文输出" in result + assert create_exec.await_args.args == ( + "pwsh.exe", + "-Command", + "utf8-prefix;\nGet-Location", + ) + assert create_exec.await_args.kwargs["env"] == {"PYTHONUTF8": "1"} + create_shell.assert_not_called() + + +@pytest.mark.anyio +async def test_terminal_session_uses_selected_windows_shell() -> None: + """后台会话与一次性命令必须共享同一 Windows shell 策略。""" + shell = AgentShell( + kind="git-bash", + executable="bash.exe", + arguments=("-lc",), + ) + process = _fake_process() + create_exec = AsyncMock(return_value=process) + manager = _TerminalSessionManager() + + with ( + patch( + "app.agent.tools.impl._terminal_session.resolve_agent_shell", + return_value=shell, + ), + patch( + "app.agent.tools.impl._terminal_session.asyncio.create_subprocess_exec", + create_exec, + ), + patch( + "app.agent.tools.impl._terminal_session.asyncio.create_subprocess_shell" + ) as create_shell, + ): + session = await manager._start_pipe_session( + "git status", "C:/MoviePilot", {"PYTHONUTF8": "1"} + ) + assert session.wait_task is not None + await session.wait_task + + assert create_exec.await_args.args == ("bash.exe", "-lc", "git status") + assert create_exec.await_args.kwargs["env"] == {"PYTHONUTF8": "1"} + create_shell.assert_not_called() + + +@pytest.mark.anyio +async def test_stdio_mcp_inherits_agent_utf8_environment() -> None: + """stdio MCP 子进程也必须继承 Windows Agent 的 UTF-8 环境。""" + process = _fake_process() + create_exec = AsyncMock(return_value=process) + server = AgentMcpServerConfig( + id="fake", + name="Fake MCP", + transport="stdio", + command="server.exe", + env={"PYTHONIOENCODING": "cp936"}, + ) + session = _StdioMcpSession(server) + + with ( + patch( + "app.agent.mcp.build_agent_subprocess_env", + return_value={"PYTHONIOENCODING": "utf-8"}, + ) as build_env, + patch("app.agent.mcp.asyncio.create_subprocess_exec", create_exec), + ): + await session.__aenter__() + await session.__aexit__(None, None, None) + + build_env.assert_called_once_with(server.env) + assert create_exec.await_args.kwargs["env"] == {"PYTHONIOENCODING": "utf-8"} + + +def test_service_script_uses_windows_utf8_text_subprocess_kwargs(tmp_path: Path) -> None: + """固定 Skill Python 子进程应显式使用 UTF-8 文本解码和环境。""" + completed = subprocess.CompletedProcess( + args=[], + returncode=0, + stdout=json.dumps({"success": True}, ensure_ascii=False), + stderr="", + ) + with ( + patch( + "app.agent.tools.impl.service.get_runtime_setting", + return_value=tmp_path, + ), + patch( + "app.agent.tools.impl.service.agent_text_subprocess_kwargs", + return_value={ + "encoding": "utf-8", + "errors": "replace", + "env": {"PYTHONUTF8": "1"}, + }, + ), + patch( + "app.agent.tools.impl.service.subprocess.run", + return_value=completed, + ) as runner, + ): + payload = _run_service_script( + relative_script="skills/demo.py", + selector_flag=None, + selector_value=None, + action="list", + arguments={}, + ) + + assert payload["success"] is True + assert runner.call_args.kwargs["encoding"] == "utf-8" + assert runner.call_args.kwargs["errors"] == "replace" + assert runner.call_args.kwargs["env"] == {"PYTHONUTF8": "1"} + + +def test_audio_ffmpeg_uses_windows_utf8_text_subprocess_kwargs(tmp_path: Path) -> None: + """Agent 音频 ffmpeg 文本输出也应走统一的 Windows UTF-8 解码参数。""" + input_path = tmp_path / "input.wav" + input_path.write_bytes(b"wav") + expected_output = input_path.with_suffix(".opus") + + def run_ffmpeg(_command, **_kwargs): + """模拟 ffmpeg 成功生成输出文件。""" + expected_output.write_bytes(b"opus") + return subprocess.CompletedProcess([], 0, "", "") + + provider = OpenAIChatAudioProvider() + with ( + patch( + "app.agent.llm.capability.agent_text_subprocess_kwargs", + return_value={ + "encoding": "utf-8", + "errors": "replace", + "env": {"PYTHONUTF8": "1"}, + }, + ), + patch( + "app.agent.llm.capability.subprocess.run", + side_effect=run_ffmpeg, + ) as runner, + ): + output_path = provider._convert_wav_to_opus(input_path) + + assert output_path == expected_output + assert runner.call_args.kwargs["encoding"] == "utf-8" + assert runner.call_args.kwargs["errors"] == "replace" + assert runner.call_args.kwargs["env"] == {"PYTHONUTF8": "1"}