diff --git a/pyproject.toml b/pyproject.toml
index 790741a..3f0a725 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -17,7 +17,7 @@ license = { text = "MIT" }
 name = "pyflowx"
 readme = "README.md"
 requires-python = ">=3.8"
-version = "0.1.8"
+version = "0.1.10"
 
 [project.scripts]
 autofmt  = "pyflowx.cli.autofmt:main"
diff --git a/src/pyflowx/__init__.py b/src/pyflowx/__init__.py
index c71739b..7612115 100644
--- a/src/pyflowx/__init__.py
+++ b/src/pyflowx/__init__.py
@@ -84,7 +84,7 @@ from .runner import CliExitCode, CliRunner
 from .storage import JSONBackend, MemoryBackend, StateBackend
 from .task import TaskCmd, TaskEvent, TaskResult, TaskSpec, TaskStatus
 
-__version__ = "0.1.8"
+__version__ = "0.1.10"
 
 __all__ = [
     "IS_LINUX",
diff --git a/src/pyflowx/cli/pymake.py b/src/pyflowx/cli/pymake.py
index 4adbc12..f8d980e 100644
--- a/src/pyflowx/cli/pymake.py
+++ b/src/pyflowx/cli/pymake.py
@@ -20,13 +20,15 @@ def maturin_build_cmd() -> list[str]:
     """
     command = ["maturin", "build", "-r"].copy()
     if Constants.IS_WINDOWS:
-        command.extend([
-            "--target",
-            "x86_64-win7-windows-msvc",
-            "-Zbuild-std",
-            "-i",
-            "python3.8",
-        ])
+        command.extend(
+            [
+                "--target",
+                "x86_64-win7-windows-msvc",
+                "-Zbuild-std",
+                "-i",
+                "python3.8",
+            ]
+        )
     return command
 
 
@@ -99,26 +101,26 @@ def main():
       pymake type        # 类型检查
     """
     runner = px.CliRunner(
-        strategy="thread",
+        strategy="sequential",
         description="PyMake - Python 构建工具",
         graphs={
             # 构建命令
             "b": px.Graph.from_specs([uv_build]),
             "bc": px.Graph.from_specs([maturin_build]),
-            "ba": px.Graph.from_specs([uv_build, maturin_build]),
+            "ba": px.Graph.from_specs(["b", "bc"]),
             # 安装命令
             "sync": px.Graph.from_specs([uv_sync]),
             # 清理命令
             "c": px.Graph.from_specs([git_clean]),
             # 开发工具
-            "bump": px.Graph.from_specs([git_clean, typecheck, ruff_lint, ruff_format, bump]),
+            "bump": px.Graph.from_specs(["c", "tc", bump]),
             "cov": px.Graph.from_specs([git_clean, test_coverage]),
             "doc": px.Graph.from_specs([doc]),
             "lint": px.Graph.from_specs([ruff_lint, ruff_format]),
             "pb": px.Graph.from_specs([twine_publish, hatch_publish]),
             "t": px.Graph.from_specs([test]),
             "tf": px.Graph.from_specs([test_fast]),
-            "tc": px.Graph.from_specs([typecheck, ruff_lint, ruff_format]),
+            "tc": px.Graph.from_specs([typecheck, "lint"]),
             "tox": px.Graph.from_specs([tox]),
             # 发布命令
             "p": px.Graph.from_specs([git_clean, git_push, git_push_tags]),
diff --git a/src/pyflowx/graph.py b/src/pyflowx/graph.py
index 1ec0eea..e4122f8 100644
--- a/src/pyflowx/graph.py
+++ b/src/pyflowx/graph.py
@@ -57,18 +57,59 @@ class Graph:
         return self
 
     @classmethod
-    def from_specs(cls, specs: Iterable[TaskSpec[Any]]) -> Graph:
-        """从可迭代的 task spec 构建图。
+    def from_specs(cls, specs: Iterable[TaskSpec[Any] | str]) -> Graph:
+        """从可迭代的 task spec 构建图.
 
         先收集所有 spec，再统一校验。这意味着任务可以引用*后出现*的
         依赖——顺序无关，就像声明式配置文件的读取方式。
+
+        支持字符串引用，允许引用其他命令图中的任务。
+        字符串引用将在CliRunner中解析展开。
+
+        Parameters
+        ----------
+        specs : Iterable[TaskSpec[Any] | str]
+            TaskSpec对象或字符串引用的列表
+
+        Returns
+        -------
+        Graph
+            构建完成的图
+
+        Note
+        -----
+        字符串引用格式：
+        - "command_name" - 引用整个命令图
+        - "command_name.task_name" - 引用特定任务
+
+        Examples
+        --------
+        >>> graph = Graph.from_specs([
+        ...     TaskSpec("build", cmd=["uv", "build"]),
+        ...     "test",  # 引用test命令图
+        ... ])
         """
         graph = cls()
+        pending_refs: list[str] = []
+
         for spec in specs:
-            if spec.name in graph.specs:
-                raise DuplicateTaskError(spec.name)
-            graph.specs[spec.name] = spec
-            graph.deps[spec.name] = spec.depends_on
+            if isinstance(spec, str):
+                # 字符串引用，稍后解析
+                pending_refs.append(spec)
+            elif isinstance(spec, TaskSpec):
+                if spec.name in graph.specs:
+                    raise DuplicateTaskError(spec.name)
+                graph.specs[spec.name] = spec
+                graph.deps[spec.name] = spec.depends_on
+            else:
+                raise TypeError(f"from_specs只接受TaskSpec或str，收到: {type(spec)}")
+
+        # 存储待解析的引用
+        if pending_refs:
+            # 使用特殊属性存储引用，稍后在CliRunner中解析
+            # 由于Graph是frozen dataclass，我们需要特殊处理
+            object.__setattr__(graph, "_pending_refs", pending_refs)
+
         graph._validate_references()
         graph.validate()
         return graph
diff --git a/src/pyflowx/runner.py b/src/pyflowx/runner.py
index 8b645bf..80a1b67 100644
--- a/src/pyflowx/runner.py
+++ b/src/pyflowx/runner.py
@@ -114,6 +114,148 @@ class CliRunner:
         if not self.graphs:
             raise ValueError("CliRunner 至少需要一个命令 (通过关键字参数提供)")
 
+        # 解析并展开字符串引用
+        self._resolve_graph_refs()
+
+    def _resolve_graph_refs(self) -> None:
+        """解析并展开图中的字符串引用.
+
+        支持两种引用格式：
+        1. "command_name" - 引用整个命令图
+        2. "command_name.task_name" - 引用特定任务
+
+        递归解析所有引用，直到所有图都只包含TaskSpec对象。
+        """
+        resolved_graphs: dict[str, Graph] = {}
+
+        for cmd_name, graph in self.graphs.items():
+            resolved_graph = self._expand_refs(graph, cmd_name)
+            resolved_graphs[cmd_name] = resolved_graph
+
+        # 更新graphs字典
+        object.__setattr__(self, "graphs", resolved_graphs)
+
+    def _expand_refs(self, graph: Graph, current_cmd: str) -> Graph:
+        """展开图中的字符串引用.
+
+        Parameters
+        ----------
+        graph : Graph
+            包含可能的字符串引用的图
+        current_cmd : str
+            当前命令名（用于避免循环引用）
+
+        Returns
+        -------
+        Graph
+            展开后的图，只包含TaskSpec对象
+
+        Note
+        -----
+        引用按顺序展开，后续引用的任务依赖于前面引用的任务完成。
+        例如：["c", "tc", bump] 会展开为：
+        - c的所有任务（无依赖）
+        - tc的所有任务（依赖于c的最后一个任务）
+        - bump任务（依赖于tc的最后一个任务）
+        """
+        # 检查是否有待解析的引用
+        pending_refs = getattr(graph, "_pending_refs", None)
+        if not pending_refs:
+            return graph
+
+        # 收集所有TaskSpec（包括原始图中的）
+        all_specs: list[TaskSpec[Any]] = []
+        for spec in graph.all_specs().values():
+            all_specs.append(spec)
+
+        # 记录每个引用展开后的所有任务名，用于建立依赖链
+        previous_ref_last_task: str | None = None
+
+        # 解析每个引用，并建立依赖关系
+        for ref in pending_refs:
+            expanded_specs = self._parse_ref(ref, current_cmd)
+
+            # 如果有前面的引用，让当前引用的所有任务依赖于前面引用的最后一个任务
+            if previous_ref_last_task and expanded_specs:
+                # 为当前引用的每个任务添加依赖
+                for i, task in enumerate(expanded_specs):
+                    # 只为没有依赖的任务添加依赖，或者为第一个任务添加依赖
+                    if i == 0 or not task.depends_on:
+                        updated_task = replace(task, depends_on=tuple({*task.depends_on, previous_ref_last_task}))
+                        expanded_specs[i] = updated_task
+
+            # 记录当前引用的最后一个任务名
+            if expanded_specs:
+                previous_ref_last_task = expanded_specs[-1].name
+
+            all_specs.extend(expanded_specs)
+
+        # 如果原始图中有TaskSpec，让它们依赖于最后一个引用的任务
+        original_specs = list(graph.all_specs().values())
+        if previous_ref_last_task and original_specs:
+            # 为每个原始TaskSpec添加依赖
+            for i, original_task in enumerate(original_specs):
+                # 只为第一个原始任务添加依赖，或者为没有依赖的任务添加依赖
+                if i == 0 or not original_task.depends_on:
+                    updated_task = replace(
+                        original_task, depends_on=tuple({*original_task.depends_on, previous_ref_last_task})
+                    )
+                    all_specs[all_specs.index(original_task)] = updated_task
+
+        # 创建新的图（不包含引用）
+        return Graph.from_specs(all_specs)
+
+    def _parse_ref(self, ref: str, current_cmd: str) -> list[TaskSpec[Any]]:
+        """解析单个字符串引用.
+
+        Parameters
+        ----------
+        ref : str
+            引用字符串（如"tc"或"tc.lint"）
+        current_cmd : str
+            当前命令名（用于避免循环引用）
+
+        Returns
+        -------
+        list[TaskSpec[Any]]
+            解析后的TaskSpec列表
+
+        Raises
+        ------
+        ValueError
+            如果引用无效或存在循环引用
+        """
+        # 避免循环引用
+        if ref == current_cmd:
+            raise ValueError(f"循环引用: 命令 '{current_cmd}' 引用了自己")
+
+        # 解析引用格式
+        if "." in ref:
+            # 特定任务引用: "command_name.task_name"
+            cmd_name, task_name = ref.split(".", 1)
+            if cmd_name not in self.graphs:
+                raise ValueError(f"引用的命令 '{cmd_name}' 不存在")
+
+            # 获取特定任务
+            ref_graph = self.graphs[cmd_name]
+            if task_name not in ref_graph.all_specs():
+                raise ValueError(f"任务 '{task_name}' 不存在于命令 '{cmd_name}' 中")
+
+            return [ref_graph.all_specs()[task_name]]
+        else:
+            # 整个命令图引用: "command_name"
+            cmd_name = ref
+            if cmd_name not in self.graphs:
+                raise ValueError(f"引用的命令 '{cmd_name}' 不存在")
+
+            # 获取整个图的所有任务
+            ref_graph = self.graphs[cmd_name]
+
+            # 递归展开引用（如果引用的图也有引用）
+            ref_graph = self._expand_refs(ref_graph, cmd_name)
+
+            return list(ref_graph.all_specs().values())
+
     # ------------------------------------------------------------------ #
     # 内省
     # ------------------------------------------------------------------ #
diff --git a/tests/cli/test_packtool.py b/tests/cli/test_packtool.py
index e1a4549..c4bb288 100644
--- a/tests/cli/test_packtool.py
+++ b/tests/cli/test_packtool.py
@@ -90,16 +90,14 @@ class TestInstallEmbedPython:
         output_dir = tmp_path / "python"
 
         # Create a mock cache file that doesn't exist (force download)
-        with patch("urllib.request.urlretrieve") as mock_urlretrieve, \
-             patch("zipfile.ZipFile") as mock_zipfile:
-
+        with patch("urllib.request.urlretrieve") as mock_urlretrieve, patch("zipfile.ZipFile") as mock_zipfile:
             # Mock successful download
             mock_urlretrieve.return_value = None
             mock_zip_instance = MagicMock()
             mock_zipfile.return_value.__enter__.return_value = mock_zip_instance
 
             # Ensure cache doesn't exist by using tmp_path as cache dir
-            with patch.object(packtool, 'DEFAULT_CACHE_DIR', str(tmp_path / ".cache")):
+            with patch.object(packtool, "DEFAULT_CACHE_DIR", str(tmp_path / ".cache")):
                 packtool.install_embed_python("3.10", output_dir)
 
                 # Verify download was called
diff --git a/tests/test_command_refs.py b/tests/test_command_refs.py
new file mode 100644
index 0000000..7af3c50
--- /dev/null
+++ b/tests/test_command_refs.py
@@ -0,0 +1,196 @@
+"""Tests for command reference feature in CliRunner."""
+
+from __future__ import annotations
+
+import pytest
+
+import pyflowx as px
+
+
+class TestCommandReferences:
+    """Test string references in Graph.from_specs."""
+
+    def test_simple_command_reference(self) -> None:
+        """Should expand simple command reference."""
+        build_task = px.TaskSpec("build", cmd=["echo", "building"])
+        test_task = px.TaskSpec("test", cmd=["echo", "testing"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "build": px.Graph.from_specs([build_task]),
+                "test": px.Graph.from_specs([test_task]),
+                "all": px.Graph.from_specs([build_task, "test"]),
+            },
+        )
+
+        # Check that 'all' command has both tasks
+        all_tasks = list(runner.graphs["all"].all_specs().keys())
+        assert "build" in all_tasks
+        assert "test" in all_tasks
+        assert len(all_tasks) == 2
+
+    def test_multiple_command_references(self) -> None:
+        """Should expand multiple command references."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+        task2 = px.TaskSpec("task2", cmd=["echo", "2"])
+        task3 = px.TaskSpec("task3", cmd=["echo", "3"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "cmd1": px.Graph.from_specs([task1]),
+                "cmd2": px.Graph.from_specs([task2]),
+                "cmd3": px.Graph.from_specs([task3]),
+                "all": px.Graph.from_specs(["cmd1", "cmd2", "cmd3"]),
+            },
+        )
+
+        # Check that 'all' command has all tasks
+        all_tasks = list(runner.graphs["all"].all_specs().keys())
+        assert set(all_tasks) == {"task1", "task2", "task3"}
+
+    def test_specific_task_reference(self) -> None:
+        """Should expand specific task reference."""
+        lint_task = px.TaskSpec("lint", cmd=["echo", "linting"])
+        format_task = px.TaskSpec("format", cmd=["echo", "formatting"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "lint": px.Graph.from_specs([lint_task, format_task]),
+                "quick": px.Graph.from_specs(["lint.lint"]),
+            },
+        )
+
+        # Check that 'quick' command only has lint task
+        quick_tasks = list(runner.graphs["quick"].all_specs().keys())
+        assert quick_tasks == ["lint"]
+
+    def test_nested_command_reference(self) -> None:
+        """Should expand nested command references."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+        task2 = px.TaskSpec("task2", cmd=["echo", "2"])
+        task3 = px.TaskSpec("task3", cmd=["echo", "3"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "cmd1": px.Graph.from_specs([task1]),
+                "cmd2": px.Graph.from_specs(["cmd1", task2]),
+                "cmd3": px.Graph.from_specs(["cmd2", task3]),
+            },
+        )
+
+        # Check that 'cmd3' has all tasks
+        cmd3_tasks = list(runner.graphs["cmd3"].all_specs().keys())
+        assert set(cmd3_tasks) == {"task1", "task2", "task3"}
+
+    def test_circular_reference_error(self) -> None:
+        """Should raise error for circular references."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+
+        with pytest.raises(ValueError, match="循环引用"):
+            px.CliRunner(
+                strategy="sequential",
+                graphs={
+                    "cmd1": px.Graph.from_specs(["cmd1", task1]),
+                },
+            )
+
+    def test_invalid_command_reference_error(self) -> None:
+        """Should raise error for invalid command reference."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+
+        with pytest.raises(ValueError, match="引用的命令 'invalid' 不存在"):
+            px.CliRunner(
+                strategy="sequential",
+                graphs={
+                    "cmd1": px.Graph.from_specs(["invalid", task1]),
+                },
+            )
+
+    def test_invalid_task_reference_error(self) -> None:
+        """Should raise error for invalid task reference."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+
+        with pytest.raises(ValueError, match="任务 'invalid' 不存在于命令 'cmd1' 中"):
+            px.CliRunner(
+                strategy="sequential",
+                graphs={
+                    "cmd1": px.Graph.from_specs([task1]),
+                    "cmd2": px.Graph.from_specs(["cmd1.invalid"]),
+                },
+            )
+
+    def test_reference_preserves_dependencies(self) -> None:
+        """Should preserve dependencies when expanding references."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+        task2 = px.TaskSpec("task2", cmd=["echo", "2"], depends_on=("task1",))
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "cmd1": px.Graph.from_specs([task1, task2]),
+                "cmd2": px.Graph.from_specs(["cmd1"]),
+            },
+        )
+
+        # Check that dependencies are preserved
+        cmd2_deps = runner.graphs["cmd2"].deps
+        assert cmd2_deps["task2"] == ("task1",)
+
+    def test_mixed_references_and_tasks(self) -> None:
+        """Should handle mixed references and direct tasks."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "1"])
+        task2 = px.TaskSpec("task2", cmd=["echo", "2"])
+        task3 = px.TaskSpec("task3", cmd=["echo", "3"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "cmd1": px.Graph.from_specs([task1, task2]),
+                "cmd2": px.Graph.from_specs(["cmd1", task3]),
+            },
+        )
+
+        # Check that 'cmd2' has all tasks
+        cmd2_tasks = list(runner.graphs["cmd2"].all_specs().keys())
+        assert set(cmd2_tasks) == {"task1", "task2", "task3"}
+
+    def test_execution_order_with_references(self) -> None:
+        """Should execute references in correct order."""
+        task1 = px.TaskSpec("task1", cmd=["echo", "step1"])
+        task2 = px.TaskSpec("task2", cmd=["echo", "step2"])
+        task3 = px.TaskSpec("task3", cmd=["echo", "step3"])
+        task4 = px.TaskSpec("task4", cmd=["echo", "step4"])
+        task5 = px.TaskSpec("task5", cmd=["echo", "step5"])
+
+        runner = px.CliRunner(
+            strategy="sequential",
+            graphs={
+                "cmd1": px.Graph.from_specs([task1]),
+                "cmd2": px.Graph.from_specs([task2, task3]),
+                "cmd3": px.Graph.from_specs([task4]),
+                "ordered": px.Graph.from_specs(["cmd1", "cmd2", "cmd3", task5]),
+            },
+        )
+
+        # Check execution order through layers
+        layers = runner.graphs["ordered"].layers()
+
+        # Layer 1 should have task1 (cmd1)
+        assert "task1" in layers[0]
+
+        # Layer 2 should have task2 and task3 (cmd2)
+        assert "task2" in layers[1]
+        assert "task3" in layers[1]
+
+        # Layer 3 should have task4 (cmd3)
+        assert "task4" in layers[2]
+
+        # Layer 4 should have task5 (original task)
+        assert "task5" in layers[3]
+
+        # Verify total layers
+        assert len(layers) == 4