diff --git a/docs4/reme_design.md b/docs4/reme_design.md
index 15e7ee28..fab618ff 100644
--- a/docs4/reme_design.md
+++ b/docs4/reme_design.md
@@ -42,6 +42,7 @@ reme4 version
 | 🔎 search  | 🔍 `search` (`search_step`)                      | `call_server("search", query=…, …)`                         | 📥 `query:str` ⭐ ｜ 🎚️ `limit:int=5`(>0) ｜ 🎚️ `min_score:float=0.0` ｜ ⚖️ `vector_weight:float=0.7` ∈[0,1]（keyword 权 = 1-vw）｜ 🔀 `candidate_multiplier:float=3.0`（candidates = min(200, limit×mult)）｜ 🔗 `expand_links:bool=True` ｜ 🔢 `max_links_per_direction:int=10` ｜ 🎚️ `search_filter:dict={}` ｜ 📤 `answer` 每命中一行 `path:start-end [score=… vector=… keyword=…] text` + 缩进的 `→ outlinks (n)` / `← inlinks (n)` + `via predicate=… anchor=#…` ｜ 📊 `metadata.results` / `metadata.link_expansion` / `metadata.counts={vector,keyword,returned,hybrid}` ｜ 🛠️ 并行 `vector_search` + `keyword_search` → RRF 融合（K=60，按 chunk.id 合并）→ `min_score` 过滤 → `limit` 截断 → 邻居 meta 注入 |
 | 🧪 demo    | 🪄 `demo_echo` (`demo_echo_step1` + `step2`)     | `call_server("demo_echo", query=…, min_score=…)`            | 📥 `query:str=""` ｜ 🎚️ `min_score:float=0.5` ｜ 🛠️ step1：`processed_query = query.strip().lower()`，`adjusted_min_score = min_score * 0.9`，写回 context ｜ 📤 step2：`answer = "echo: {processed_query} (min_score={adjusted_min_score})"` ｜ 📊 `metadata = {step, query, min_score, processed_query, adjusted_min_score}`                                                                                                                                                                                                                                                                                                                                                          |
 | 🌊 demo    | 🌊 `stream_demo` (`stream_demo_step1` + `step2`) | `call_server("stream_demo", query=…, repeat=…, interval=…)` | 📥 `query:str=""` ｜ 🎚️ `repeat:int=10` ｜ 🎚️ `interval:float=0.1`（秒/字符）｜ 🛠️ step1：`stream_text = query * repeat` 写回 context ｜ 📤 step2：按字符 `add_stream_string(ch, ChunkEnum.CONTENT)` 流式输出，`asyncio.sleep(interval)` 节流                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
+| 📂 crud    | 📖 `read` (`read_step`)                          | `call_server("read", path=…, …)`                            | 📥 `path:str` ⭐（**完整相对路径**，相对于 `working_dir`；绝对路径会被拒绝；非 `.md` 后缀拒绝）｜ 🎚️ `start_line:int=null`（1-based, 含端点）｜ 🎚️ `end_line:int=null`（1-based, 含端点）｜ 🎚️ `max_bytes:int=51200`（截断阈值）｜ 📤 `answer = 选中的行内容`，超过 `max_bytes` 时附加 `--- TRUNCATED ---` 续读指引（`start_line=…`）｜ 📊 `metadata.path` / `metadata.total_lines`（出错路径才会附带）｜ 🛠️ 流程：`BaseStep.resolve_path(raw, require_md=True)` → `aiofiles.os.stat` → `read_file_safe`（utf-8-sig BOM 容忍、UnicodeDecodeError fallback `errors=ignore`）→ `split("\n")` 切片 `[s-1:e]` → `truncate_text_output` 按字节截断保行                                                                                              |
 
 使用示例：
 
@@ -66,6 +67,11 @@ reme4 version
 reme4 reindex
 reme4 search query="latency 问题" limit=10 min_score=0.2 vector_weight=0.6
 
+# 读取 working_dir 下的 markdown（完整相对路径；无后缀自动补 .md；可按行切片或限制字节）
+reme4 read path=Templates/Recipe.md
+reme4 read path=Notes start_line=1 end_line=20
+reme4 read path=Big.md max_bytes=4096
+
 # 通过 MCP backend 调用
 reme4 search query="..." backend=mcp
 ```
diff --git a/reme4/components/job/base_job.py b/reme4/components/job/base_job.py
index e4681bf1..2583989f 100644
--- a/reme4/components/job/base_job.py
+++ b/reme4/components/job/base_job.py
@@ -49,5 +49,6 @@ class BaseJob(BaseComponent):
                 await step(context)
         except Exception as e:
             self.logger.exception(f"Failed to execute job: {e}")
+            context.response.success = False
             context.response.answer = str(e)
         return context.response
diff --git a/reme4/config/default.yaml b/reme4/config/default.yaml
index 1a501912..6e543ec9 100644
--- a/reme4/config/default.yaml
+++ b/reme4/config/default.yaml
@@ -96,6 +96,26 @@ jobs:
     steps:
       - backend: search_step
 
+  - backend: base
+    name: read
+    description: "read a markdown file (relative path under working_dir)"
+    parameters:
+      type: object
+      properties:
+        path:
+          type: string
+          description: "relative path under the working_dir (no absolute paths); markdown only"
+        start_line:
+          type: integer
+          description: "Optional, first line to read (1-based, inclusive)"
+        end_line:
+          type: integer
+          description: "Optional, last line to read (1-based, inclusive)"
+      required:
+        - path
+    steps:
+      - backend: read_step
+
   - backend: stream
     name: stream_demo
     description: "stream demo job: repeat query 10x and stream char-by-char"
diff --git a/reme4/constants.py b/reme4/constants.py
index fce66e64..819fbadd 100644
--- a/reme4/constants.py
+++ b/reme4/constants.py
@@ -5,3 +5,8 @@ REME_SERVICE_INFO = "REME_SERVICE_INFO"
 REME_DEFAULT_HOST = "127.0.0.1"
 
 REME_DEFAULT_PORT = 2333
+
+# CRUD steps: file IO limits and truncation marker (shared across CRUD steps).
+DEFAULT_MAX_BYTES = 50 * 1024
+MAX_FILE_READ_BYTES = 200 * 1024 * 1024
+TRUNCATION_NOTICE_MARKER = "<<TRUNCATION_NOTICE>>"
diff --git a/reme4/steps/__init__.py b/reme4/steps/__init__.py
index fa25273c..70878e5f 100644
--- a/reme4/steps/__init__.py
+++ b/reme4/steps/__init__.py
@@ -1,9 +1,11 @@
 """steps"""
 
 from . import common
+from . import crud
 from .base_step import BaseStep
 
 __all__ = [
     "common",
+    "crud",
     "BaseStep",
 ]
diff --git a/reme4/steps/base_step.py b/reme4/steps/base_step.py
index d32c4445..f8a4d7aa 100644
--- a/reme4/steps/base_step.py
+++ b/reme4/steps/base_step.py
@@ -2,6 +2,7 @@
 
 import copy
 from abc import abstractmethod, ABC
+from pathlib import Path
 from typing import TypeVar, TYPE_CHECKING
 
 from agentscope.formatter import FormatterBase
@@ -83,7 +84,20 @@ class BaseStep(ABC):
             self.context.apply_mapping(self.output_mapping)
         return result
 
-    def _resolve(self, key: str, base_cls: type[T], comp_enum: ComponentEnum, attr: str | None = None) -> T:
+    @property
+    def working_path(self) -> Path:
+        """Resolved working directory from app context or cwd."""
+        if self.app_context is None:
+            return Path.cwd()
+        return Path(self.app_context.app_config.working_dir)
+
+    def _resolve(
+        self,
+        key: str,
+        base_cls: type[T],
+        comp_enum: ComponentEnum,
+        attr: str | None = None,
+    ) -> T:
         """Return a kwargs-supplied instance, or look one up by name in the app registry."""
         # 1. Step init kwargs, 2. Runtime context (run_job kwargs), 3. App registry by name.
         for source in (self.kwargs, self.context or {}):
diff --git a/reme4/steps/crud/__init__.py b/reme4/steps/crud/__init__.py
new file mode 100644
index 00000000..de9ae9ce
--- /dev/null
+++ b/reme4/steps/crud/__init__.py
@@ -0,0 +1,7 @@
+"""CRUD steps for markdown files under the working_dir."""
+
+from .read import ReadStep
+
+__all__ = [
+    "ReadStep",
+]
diff --git a/reme4/steps/crud/_file_io.py b/reme4/steps/crud/_file_io.py
new file mode 100644
index 00000000..4beaa35d
--- /dev/null
+++ b/reme4/steps/crud/_file_io.py
@@ -0,0 +1,109 @@
+"""Shared filesystem helpers for CRUD steps (path gating, safe read, truncation)."""
+
+from pathlib import Path
+
+import aiofiles
+import aiofiles.os
+
+from ...constants import DEFAULT_MAX_BYTES, MAX_FILE_READ_BYTES, TRUNCATION_NOTICE_MARKER
+from ...utils import get_logger
+
+logger = get_logger()
+
+
+def resolve_path(working_path: Path, raw: str) -> tuple[Path | None, str | None]:
+    """Resolve a relative `path=` argument under self.working_path.
+
+    Rules:
+        - the caller supplies the full relative path under ``self.working_path``;
+        absolute paths are rejected.
+    Returns ``(abs_path, None)`` on success, or ``(None, error_message)`` on failure.
+    Filetype-specific gating (e.g. markdown-only / suffix auto-append) is
+    layered on top by callers — see ``reme4/steps/crud/_file_io.py::gate_md``.
+    """
+    if not raw or not str(raw).strip():
+        return None, "`path` is required"
+    s = str(raw).strip()
+    p = Path(s)
+    if p.is_absolute():
+        logger.info("absolute path detected, recommmending relative paths")
+        return p, None
+    return working_path / p, None
+
+
+def gate_md(target: Path, raw: str) -> tuple[Path | None, str | None]:
+    """Markdown-only gate: auto-append `.md` when no suffix; reject any non-`.md` suffix.
+
+    Layered on top of ``BaseStep.resolve_path`` to keep filetype-specific rules
+    out of the generic path resolver.
+    """
+    if target.suffix == "":
+        return target.with_suffix(".md"), None
+    if target.suffix.lower() != ".md":
+        return None, (f"path {raw!r} is not a markdown file; this command only supports .md files")
+    return target, None
+
+
+async def read_file_safe(file_path, max_bytes: int = MAX_FILE_READ_BYTES) -> str:
+    """Read file with utf-8-sig (BOM-tolerant), fallback to errors='ignore'."""
+    stat = await aiofiles.os.stat(str(file_path))
+    read_size = min(stat.st_size, max_bytes)
+    try:
+        async with aiofiles.open(str(file_path), "r", encoding="utf-8-sig") as f:
+            return await f.read(read_size)
+    except UnicodeDecodeError:
+        async with aiofiles.open(
+            str(file_path),
+            "r",
+            encoding="utf-8-sig",
+            errors="ignore",
+        ) as f:
+            return await f.read(read_size)
+
+
+def truncate_text_output(
+    text: str,
+    *,
+    start_line: int = 1,
+    total_lines: int = 0,
+    max_bytes: int = DEFAULT_MAX_BYTES,
+    file_path: str | None = None,
+    encoding: str = "utf-8",
+) -> str:
+    """Truncate text by bytes preserving line integrity; append a continuation notice.
+
+    See qwenpaw `tools/utils.py` for the same semantics. Returns text unchanged when
+    it fits within max_bytes, when max_bytes <= 0, or when the last line itself
+    exceeds max_bytes (unhandled edge case).
+    """
+    if not text or max_bytes <= 0:
+        return text
+
+    try:
+        text_bytes = text.encode(encoding)
+        if len(text_bytes) <= max_bytes:
+            return text
+
+        truncated = text_bytes[:max_bytes]
+        result = truncated.decode(encoding, errors="ignore")
+        newline_count = result.count("\n")
+        next_line = start_line + max(1, newline_count)
+
+        if next_line <= total_lines:
+            read_from = next_line
+        elif start_line < total_lines:
+            read_from = total_lines
+        else:
+            return result
+
+        notice = (
+            TRUNCATION_NOTICE_MARKER + f"\nThe output above was truncated."
+            f"\nThe full content is saved to the file and contains {total_lines} lines in total."
+            f"\nThis excerpt starts at line {start_line} and covers the next {max_bytes} bytes."
+            f"\nIf the current content is not enough, call `read` with file={file_path or ''} "
+            f"start_line={read_from} to read more."
+        )
+        return result + notice
+    except Exception:
+        logger.warning("truncate_text_output failed, returning original text", exc_info=True)
+        return text
diff --git a/reme4/steps/crud/read.py b/reme4/steps/crud/read.py
new file mode 100644
index 00000000..55285efa
--- /dev/null
+++ b/reme4/steps/crud/read.py
@@ -0,0 +1,91 @@
+"""Read a markdown file from the vault, with line-range slicing and byte-truncation."""
+
+from ._file_io import (
+    gate_md,
+    resolve_path,
+    read_file_safe,
+    truncate_text_output,
+)
+from ..base_step import BaseStep
+from ...components import R
+
+
+@R.register("read_step")
+class ReadStep(BaseStep):
+    """Read a markdown file. Optional `start_line`/`end_line` for ranged reads."""
+
+    def _fail(self, message: str, **meta) -> None:
+        assert self.context is not None
+        self.context.response.success = False
+        self.context.response.answer = f"Error: {message}"
+        if meta:
+            self.context.response.metadata.update(meta)
+
+    async def execute(self):  # pylint: disable=too-many-return-statements
+        assert self.context is not None
+        raw = str(self.context.get("path") or "")
+        start_line = self.context.get("start_line")
+        end_line = self.context.get("end_line")
+
+        target, err = resolve_path(self.working_path, raw)
+        if err:
+            self._fail(err)
+            return None
+
+        target, err = gate_md(target, raw)
+        if err:
+            self._fail(err)
+            return None
+
+        for label, value in (("start_line", start_line), ("end_line", end_line)):
+            if value is None:
+                continue
+            try:
+                int(value)
+            except (TypeError, ValueError):
+                self._fail(f"{label} must be an integer, got {value!r}")
+                return None
+
+        if not target.exists():
+            self._fail(f"file {target} does not exist", path=str(target))
+            return None
+        if not target.is_file():
+            self._fail(f"path {target} is not a file", path=str(target))
+            return None
+
+        try:
+            content = await read_file_safe(target)
+        except Exception as e:
+            self._fail(f"read failed: {e}", path=str(target))
+            return None
+
+        all_lines = content.split("\n")
+        total = len(all_lines)
+        s = max(1, int(start_line) if start_line is not None else 1)
+        e = min(total, int(end_line) if end_line is not None else total)
+
+        if s > total:
+            self._fail(
+                f"start_line {s} exceeds file length ({total} lines)",
+                path=str(target),
+                total_lines=total,
+            )
+            return None
+        if s > e:
+            self._fail(f"start_line ({s}) > end_line ({e})", path=str(target))
+            return None
+
+        selected = "\n".join(all_lines[s - 1 : e])
+        text = truncate_text_output(
+            selected,
+            start_line=s,
+            total_lines=total,
+            file_path=str(target),
+        )
+
+        self.context.response.success = True
+        self.context.response.answer = text
+        self.logger.info(
+            f"[{self.name}] read path={target} lines={s}-{e}/{total} bytes={len(text.encode('utf-8'))}",
+        )
+        return self.context.response
diff --git a/tests4/unittest/test_crud_md_steps.py b/tests4/unittest/test_crud_md_steps.py
new file mode 100644
index 00000000..d60c72f9
--- /dev/null
+++ b/tests4/unittest/test_crud_md_steps.py
@@ -0,0 +1,371 @@
+"""End-to-end tests for reme4 crud_md steps: spawn `reme4 start`, drive via HTTP,
+verify responses, then shut down. Each test uses an isolated cwd so the working_dir
+(.reme by default) does not collide.
+
+CLI rule: `path=` is relative-only, rooted at the reme working_dir. A bare path with
+no suffix auto-appends `.md`; non-`.md` suffix is rejected. Absolute paths are
+rejected.
+"""
+
+import asyncio
+import os
+import tempfile
+import warnings
+from pathlib import Path
+
+from reme4.utils import call_action, call_and_check, mock_reme_server
+
+warnings.filterwarnings("ignore", category=DeprecationWarning, module="jieba")
+warnings.filterwarnings("ignore", category=DeprecationWarning, module="pkg_resources")
+
+
+class _temp_chdir:
+    """chdir to path for the duration of the block; restore on exit."""
+
+    def __init__(self, path):
+        self.path = path
+        self._old = None
+
+    def __enter__(self):
+        self._old = os.getcwd()
+        os.chdir(self.path)
+        return self
+
+    def __exit__(self, *exc):
+        os.chdir(self._old)
+
+
+def _run(coro):
+    """Run an async coroutine on a fresh isolated event loop."""
+    asyncio.run(coro)
+
+
+def _seed_md(working_dir: Path, rel: str, body: str) -> Path:
+    target = working_dir / rel
+    target.parent.mkdir(parents=True, exist_ok=True)
+    target.write_text(body, encoding="utf-8")
+    return target
+
+
+# ---------------------------------------------------------------------------
+# Individual job tests
+# ---------------------------------------------------------------------------
+
+
+def test_read_relative_path():
+    """`reme4 read path=Templates/Recipe.md` returns the file body from .reme/."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            body = "# Recipe\n\nMix flour and water.\n"
+            _seed_md(working, "Templates/Recipe.md", body)
+            async with mock_reme_server() as (host, port):
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Templates/Recipe.md",
+                    validator=lambda r: (
+                        isinstance(r, dict)
+                        and r.get("success") is True
+                        and "# Recipe" in str(r.get("answer", ""))
+                        and "flour and water" in str(r.get("answer", ""))
+                    ),
+                )
+        print("✓ test_read_relative_path passed")
+
+    _run(run())
+
+
+def test_read_no_suffix_autoappends_md():
+    """A bare path with no suffix auto-appends `.md`."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            _seed_md(working, "Templates/Recipe.md", "auto-md\n")
+            async with mock_reme_server() as (host, port):
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Templates/Recipe",
+                    validator=lambda r: (
+                        isinstance(r, dict) and r.get("success") is True and "auto-md" in str(r.get("answer", ""))
+                    ),
+                )
+        print("✓ test_read_no_suffix_autoappends_md passed")
+
+    _run(run())
+
+
+def test_read_line_range():
+    """start_line / end_line slice the file 1-based, inclusive."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            _seed_md(working, "Notes.md", "L1\nL2\nL3\nL4\nL5\n")
+            async with mock_reme_server() as (host, port):
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Notes.md",
+                    start_line=2,
+                    end_line=4,
+                    validator=lambda r: (
+                        isinstance(r, dict)
+                        and r.get("success") is True
+                        and "L2" in str(r["answer"])
+                        and "L3" in str(r["answer"])
+                        and "L4" in str(r["answer"])
+                        and "L1" not in str(r["answer"])
+                        and "L5" not in str(r["answer"])
+                    ),
+                )
+        print("✓ test_read_line_range passed")
+
+    _run(run())
+
+
+def test_read_absolute_path_rejected():
+    """Absolute paths are rejected (relative-only after refactor)."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            target = _seed_md(working, "Abs.md", "x\n")
+            async with mock_reme_server() as (host, port):
+                result = await call_action(
+                    "read",
+                    host=host,
+                    port=port,
+                    path=str(target.resolve()),
+                )
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "absolute" in str(result.get("answer", "")).lower()
+                ):
+                    raise AssertionError(f"expected absolute-path rejection, got {result!r}")
+        print("✓ test_read_absolute_path_rejected passed")
+
+    _run(run())
+
+
+def test_read_non_md_rejected():
+    """Paths whose suffix is not `.md` are rejected."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            async with mock_reme_server() as (host, port):
+                result = await call_action(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="data/foo.txt",
+                )
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "markdown" in str(result.get("answer", "")).lower()
+                ):
+                    raise AssertionError(f"expected markdown-only rejection, got {result!r}")
+        print("✓ test_read_non_md_rejected passed")
+
+    _run(run())
+
+
+def test_read_missing_file():
+    """Reading a non-existent file should fail with a clear error."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            async with mock_reme_server() as (host, port):
+                result = await call_action(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="NotThere.md",
+                )
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "does not exist" in str(result.get("answer", "")).lower()
+                ):
+                    raise AssertionError(f"expected missing-file rejection, got {result!r}")
+        print("✓ test_read_missing_file passed")
+
+    _run(run())
+
+
+def test_read_start_after_end():
+    """start_line > end_line is invalid."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            _seed_md(working, "Range.md", "a\nb\nc\n")
+            async with mock_reme_server() as (host, port):
+                result = await call_action(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Range.md",
+                    start_line=3,
+                    end_line=1,
+                )
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "start_line" in str(result.get("answer", ""))
+                ):
+                    raise AssertionError(f"expected start>end rejection, got {result!r}")
+        print("✓ test_read_start_after_end passed")
+
+    _run(run())
+
+
+def test_read_start_line_exceeds_total():
+    """start_line beyond total line count is invalid."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            _seed_md(working, "Short.md", "only-one-line\n")
+            async with mock_reme_server() as (host, port):
+                result = await call_action(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Short.md",
+                    start_line=99,
+                )
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "exceeds" in str(result.get("answer", "")).lower()
+                ):
+                    raise AssertionError(f"expected exceeds-length rejection, got {result!r}")
+        print("✓ test_read_start_line_exceeds_total passed")
+
+    _run(run())
+
+
+def test_read_truncation():
+    """A small max_bytes triggers truncation with a continuation notice."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            body = "\n".join(f"line {i}" for i in range(200)) + "\n"
+            _seed_md(working, "Big.md", body)
+            async with mock_reme_server() as (host, port):
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Big.md",
+                    max_bytes=64,
+                    validator=lambda r: (
+                        isinstance(r, dict)
+                        and r.get("success") is True
+                        and "truncated" in str(r["answer"])
+                        and "start_line=" in str(r["answer"])
+                    ),
+                )
+        print("✓ test_read_truncation passed")
+
+    _run(run())
+
+
+def test_read_empty_path_rejected():
+    """An empty `path` should be rejected."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            async with mock_reme_server() as (host, port):
+                result = await call_action("read", host=host, port=port, path="")
+                if not (
+                    isinstance(result, dict)
+                    and result.get("success") is False
+                    and "required" in str(result.get("answer", "")).lower()
+                ):
+                    raise AssertionError(f"expected `path` required rejection, got {result!r}")
+        print("✓ test_read_empty_path_rejected passed")
+
+    _run(run())
+
+
+# ---------------------------------------------------------------------------
+# Aggregate test: reuse one server instance for all read cases (faster).
+# ---------------------------------------------------------------------------
+
+
+def test_all_read_cases_one_server():
+    """Run multiple read scenarios against a single shared server for efficiency."""
+
+    async def run():
+        with tempfile.TemporaryDirectory() as tmp, _temp_chdir(tmp):
+            working = Path(tmp) / ".reme"
+            working.mkdir(parents=True, exist_ok=True)
+            _seed_md(working, "Templates/Recipe.md", "# Recipe\nbody\n")
+            _seed_md(working, "Notes.md", "L1\nL2\nL3\n")
+            async with mock_reme_server() as (host, port):
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Templates/Recipe.md",
+                    validator=lambda r: r.get("success") is True and "# Recipe" in r["answer"],
+                )
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Notes",
+                    validator=lambda r: r.get("success") is True and "L1" in r["answer"],
+                )
+                await call_and_check(
+                    "read",
+                    host=host,
+                    port=port,
+                    path="Notes.md",
+                    start_line=2,
+                    end_line=2,
+                    validator=lambda r: r.get("success") is True and r["answer"].strip() == "L2",
+                )
+        print("✓ test_all_read_cases_one_server passed")
+
+    _run(run())
+
+
+if __name__ == "__main__":
+    print("\n=== reme4 crud_md (read) E2E tests ===")
+    test_read_relative_path()
+    test_read_no_suffix_autoappends_md()
+    test_read_line_range()
+    test_read_absolute_path_rejected()
+    test_read_non_md_rejected()
+    test_read_missing_file()
+    test_read_start_after_end()
+    test_read_start_line_exceeds_total()
+    test_read_truncation()
+    test_read_empty_path_rejected()
+    test_all_read_cases_one_server()
+    print("\n所有测试通过!")