feat(benchmark): extract BEAM and LongMemEval into standalone plugins (#512)

* Simplify project implementation * Centralize benchmark agentic answer base class * Remove bundled ReMe source snapshots * Preserve local benchmark configs and plugin discovery behavior * docs: enrich job parameter descriptions in beam and lme plugin configs * Simplify project structure and remove obsolete code * Move benchmark search step configuration into BEAM and LME plugins * Rename benchmark judge packages to avoid import collisions * Remove explicit plugin package loading in favor of entry-point discovery * Export benchmark plugin Steps from public packages
2026-09-05 08:06:15 +00:00 · 2026-09-03 14:06:39 +08:00 · 2026-09-03 14:06:39 +08:00 · 3f2eb6235f
commit 3f2eb6235f
parent 8c4898999d
53 changed files with 1396 additions and 1002 deletions
--- a/.github/workflows/ci-python-quality.yml
+++ b/.github/workflows/ci-python-quality.yml
@ -34,7 +34,7 @@ jobs:
      - name: Install
        run: |
          pip install -q -e reme_studio -e ".[dev,core]"
-          pip install -q --no-deps -e plugins/auto-fin -e plugins/daily_paper
+          pip install -q --no-deps -e plugins/auto-fin -e plugins/daily_paper -e plugins/lme -e plugins/beam

      - name: Pre-commit starts
        run: pre-commit run --all-files
--- a/.github/workflows/ci-python-tests.yml
+++ b/.github/workflows/ci-python-tests.yml
@ -40,11 +40,12 @@ jobs:
          pip install -e reme_studio -e ".[dev,core]"
          pip install --no-deps -e plugins/auto-fin
          pip install -e plugins/daily_paper
+          pip install -e plugins/lme -e plugins/beam
          pip install coverage

      - name: Run unit tests
        run: |
-          coverage run -m pytest tests/unit plugins/auto-fin plugins/daily_paper \
+          coverage run -m pytest tests/unit plugins/auto-fin plugins/daily_paper plugins/lme plugins/beam \
            -v \
            --tb=long \
            -s \
--- a/AGENTS.md
+++ b/AGENTS.md
@ -46,8 +46,8 @@ and concise documentation together.
 - `reme/components/service/`: local CLI, HTTP, and MCP service backends.
 - `reme/components/`: agent wrappers, model adapters, stores, catalogs, graphs, indexes, clients, tokenizers, and
  outbound proxies.
- `reme/steps/`: registered job steps grouped by common, file I/O, index, evolve, cookbook, benchmark, and transfer
-  concerns.
+- `reme/steps/`: registered job steps grouped by common, file I/O, index, evolve, cookbook, and transfer
+  concerns, plus shared benchmark base classes under `benchmark/`.
 - `reme/utils/`: shared utilities, including service discovery, logging, web-static resolution, session I/O, token
  accounting, and wikilink handling.
 - `tests/unit/`: primary fast, isolated validation suite.
@ -56,7 +56,7 @@ and concise documentation together.
  `@agentscope-ai/reme_studio` npm static distribution.
 - `typescript/`: the independently published `@agentscope-ai/reme` package, including the shared TypeScript client and
  DeepSeek Harness and OpenClaw adapters.
- `plugins/`: installable ReMe extensions, such as Auto Fin.
+- `plugins/`: installable ReMe extensions, including Auto Fin and LME/BEAM benchmark Steps and application presets.
 - `integrations/`: adapters that connect ReMe to external agent hosts, such as Claude Code, DSH, and Hermes Agent.
 - `skills/`: standalone skills; `reme_memory` calls ReMe, while other skills may use separate tools or direct-file
  conventions.
--- a/benchmark/beam/README.md
+++ b/benchmark/beam/README.md
@ -15,8 +15,20 @@ include abstention, contradiction resolution, event ordering, information
 extraction, instruction following, knowledge update, multi-session reasoning,
 preference following, summarization, and temporal reasoning.

-> For the shared setup (dependencies, credentials, log conventions) see the
-> [top-level benchmark README](../README.md).
+Install ReMe and the BEAM plugin in editable mode from the repository root:
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/beam --editable
+reme plugins validate beam
+```
+
+The runner explicitly enables the installed `beam` plugin and combines its defaults with
+ReMe's built-in `benchmark` preset. Editable installation keeps changes under
+[`plugins/beam`](../../plugins/beam/README.md) visible without reinstalling the plugin.
+Custom application config paths still work through `reme.config` and can use `extends: benchmark`.
+This directory continues to own the runner, evaluation settings, dataset and outputs.
+Model credentials use the environment variables declared by the shared benchmark configuration.

 ## 1. Get the Dataset

@ -59,7 +71,7 @@ python benchmark/beam/run.py --eval_only               # reuse existing workspac
 | `dataset.start_index` / `num_items` | Case pagination (`num_items` `0` = all). |
 | `dataset.workspace_root` | Per-case workspace root (`benchmark/beam/workspaces/beam`). |
 | `evaluation.num_workers` | `0` = auto, `1` = sequential, `>1` = parallel. |
-| `reme.config` | ReMe config used (`beam.yaml`). |
+| `reme.config` | ReMe config used (`benchmark`). |
 | `output.dir` | Results directory (`benchmark/beam/results`). |

 ## 5. Outputs
--- a/benchmark/beam/README_ZH.md
+++ b/benchmark/beam/README_ZH.md
@ -13,7 +13,19 @@ ordering（事件排序）、information extraction（信息抽取）、instruct
 knowledge update（知识更新）、multi-session reasoning（多会话推理）、preference following
 （偏好遵循）、summarization（摘要）与 temporal reasoning（时间推理）。

-> 公共设置（依赖、凭据、日志约定）见[总评测说明](../README_ZH.md)。
+在仓库根目录以 editable 模式安装 ReMe 和 BEAM 插件：
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/beam --editable
+reme plugins validate beam
+```
+
+runner 显式启用已安装的 `beam` 插件，并将插件默认配置与 ReMe 内置的 `benchmark` 配置组合。
+editable 安装会让 [`plugins/beam`](../../plugins/beam/README_ZH.md) 下的源码修改直接生效，无需重复安装。
+本目录继续保留评测参数、数据集及输出。自定义完整应用配置路径仍可通过 `reme.config` 指定，
+并可使用 `extends: benchmark`。
+模型凭据通过公共 benchmark 配置中声明的环境变量设置。

 ## 1. 获取数据集

@ -55,7 +67,7 @@ python benchmark/beam/run.py --eval_only               # 复用已有工作区
 | `dataset.start_index` / `num_items` | case 分页（`num_items` 为 `0` 表示全部）。 |
 | `dataset.workspace_root` | case 工作区根目录（`benchmark/beam/workspaces/beam`）。 |
 | `evaluation.num_workers` | `0` = 自动，`1` = 串行，`>1` = 并行。 |
-| `reme.config` | 使用的 ReMe 配置（`beam.yaml`）。 |
+| `reme.config` | 使用的 ReMe 配置（`benchmark`）。 |
 | `output.dir` | 结果目录（`benchmark/beam/results`）。 |

 ## 5. 输出
--- a/benchmark/beam/config.yaml
+++ b/benchmark/beam/config.yaml
@ -14,7 +14,7 @@ evaluation:
  compress_session: false               # true = compress session chunks in search_v2 (query-aware); false = no compression

 reme:
-  config: "beam.yaml"                   # reme config (in reme/config/)
+  config: "benchmark"                   # shared ReMe benchmark preset; runner enables the installed beam plugin

 output:
  dir: "benchmark/beam/results"
--- a/benchmark/beam/run.py
+++ b/benchmark/beam/run.py
@ -29,7 +29,7 @@ import yaml
 from dotenv import load_dotenv

 # Load .env from project root
-_PROJECT_ROOT = Path(__file__).parent.parent.parent
+_PROJECT_ROOT = Path(__file__).resolve().parent.parent.parent
 load_dotenv(_PROJECT_ROOT / ".env")

 # Workspace root — read from config.yaml (dataset.workspace_root)
@ -151,6 +151,22 @@ def load_eval_config(config_path: str | None = None) -> dict:
    return yaml.safe_load(raw)


+def create_reme_app(config: str = "benchmark", **overrides):
+    """Create an app with the installed BEAM plugin explicitly enabled.
+
+    Plugin discovery remains environment-based; editable installation keeps local
+    plugin source changes visible to every multiprocessing worker.
+    """
+    from reme import Application
+    from reme.config import resolve_app_config
+
+    enabled_plugins = list(overrides.pop("plugins", ()) or ())
+    if "beam" not in enabled_plugins:
+        enabled_plugins.append("beam")
+    app_config = resolve_app_config(config=config, plugins=enabled_plugins, **overrides)
+    return Application(**app_config)
+
+
 # ---------------------------------------------------------------------------
 # BEAM data loading
 # ---------------------------------------------------------------------------
@ -319,8 +335,6 @@ async def evaluate_case(eval_config: dict, case_id: str, eval_only: bool = False
    Returns:
        A results dict with all questions, answers, and judgments.
    """
-    from reme import Application
-    from reme.config import resolve_app_config

    dataset_cfg = eval_config["dataset"]
    chat_size = dataset_cfg["chat_size"]
@ -375,7 +389,7 @@ async def evaluate_case(eval_config: dict, case_id: str, eval_only: bool = False
                force_init=True,
            )

-    cfg = resolve_app_config(
+    app = create_reme_app(
        config=eval_config["reme"]["config"],
        workspace_dir=workspace_dir,
        log_to_console=output_cfg.get("log_to_console", True),
@ -383,7 +397,6 @@ async def evaluate_case(eval_config: dict, case_id: str, eval_only: bool = False
        enable_logo=False,
    )

-    app = Application(**cfg)
    await app.start()

    from reme.utils.evaluation_interface import check_agent_token_usage  # noqa: E402
--- a/benchmark/longmemeval/README.md
+++ b/benchmark/longmemeval/README.md
@ -12,8 +12,20 @@ agentic (ReAct) mode, and scores the answer with an LLM-as-judge.
 Question types include single-session (user / assistant / preference),
 multi-session reasoning, knowledge update, and temporal reasoning.

-> For the shared setup (dependencies, credentials, log conventions) see the
-> [top-level benchmark README](../README.md).
+Install ReMe and the LongMemEval plugin in editable mode from the repository root:
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/lme --editable
+reme plugins validate lme
+```
+
+The runner explicitly enables the installed `lme` plugin and combines its defaults with
+ReMe's built-in `benchmark` preset. Editable installation keeps changes under
+[`plugins/lme`](../../plugins/lme/README.md) visible without reinstalling the plugin.
+Custom application config paths still work through `reme.config` and can use `extends: benchmark`.
+This directory continues to own the runner, evaluation settings, dataset and outputs.
+Model credentials use the environment variables declared by the shared benchmark configuration.

 ## 1. Get the Dataset

@ -46,7 +58,8 @@ python benchmark/longmemeval/run.py --eval_only               # reuse existing w

 1. Load the dataset (ground truth is embedded in the data file).
 2. For each item, create an isolated workspace and ingest sessions in chronological order.
-3. Trigger `auto_dream` when consecutive sessions cross the configured hour (default 23:00).
+3. If a custom application configuration enables `auto_dream`, trigger it when sessions cross the configured hour
+   (default 23:00). The packaged preset leaves it disabled.
 4. Answer each question via agentic (ReAct) mode.
 5. Judge the answer (binary yes/no) with the `answer_judge` job and print per-type accuracy.

@ -60,7 +73,7 @@ python benchmark/longmemeval/run.py --eval_only               # reuse existing w
 | `dataset.workspace_root` | Per-item workspace root (`benchmark/longmemeval/workspaces/longmemeval-s`). |
 | `evaluation.num_workers` | `0` = auto (cpu-2), `1` = sequential, `>1` = parallel. |
 | `evaluation.filter_future_sessions` | Only ingest sessions with timestamp ≤ `question_date`. |
-| `reme.config` | ReMe config used (`lme.yaml`). |
+| `reme.config` | ReMe config used (`benchmark`). |
 | `reme.dream_trigger_hour` / `dream_scan_days` / `dream_max_units` | Dream triggering behavior. |
 | `output.dir` | Results directory (`benchmark/longmemeval/results`). |

@ -93,4 +106,4 @@ agentscope==2.0.4.post1, conda reme env, 32 workers, eval-only (reusing prebuilt
 | single-session-preference | 0.633 | 36,802 | 818 | 37,620 | 3.60 |
 | single-session-user | 0.986 | 27,433 | 359 | 27,792 | 2.60 |
 | temporal-reasoning | 0.902 | 62,674 | 985 | 63,659 | 4.97 |
-| **OVERALL** | **0.894** | **43,448** | **876** | **44,324** | **3.69** |
+| **OVERALL** | **0.894** | **43,448** | **876** | **44,324** | **3.69** |
--- a/benchmark/longmemeval/README_ZH.md
+++ b/benchmark/longmemeval/README_ZH.md
@ -8,7 +8,19 @@ LongMemEval 是一个面向**多轮多会话历史的长期记忆能力**的评

 题型包括单会话（user / assistant / preference）、多会话推理、知识更新与时间推理等。

-> 公共设置（依赖、凭据、日志约定）见[总评测说明](../README_ZH.md)。
+在仓库根目录以 editable 模式安装 ReMe 和 LongMemEval 插件：
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/lme --editable
+reme plugins validate lme
+```
+
+runner 显式启用已安装的 `lme` 插件，并将插件默认配置与 ReMe 内置的 `benchmark` 配置组合。
+editable 安装会让 [`plugins/lme`](../../plugins/lme/README_ZH.md) 下的源码修改直接生效，无需重复安装。
+本目录继续保留评测参数、数据集及输出。自定义完整应用配置路径仍可通过 `reme.config` 指定，
+并可使用 `extends: benchmark`。
+模型凭据通过公共 benchmark 配置中声明的环境变量设置。

 ## 1. 获取数据集

@ -41,7 +53,7 @@ python benchmark/longmemeval/run.py --eval_only               # 复用已有工

 1. 加载数据集（ground truth 已内嵌在数据文件中）。
 2. 为每个条目创建独立工作区，按时间顺序摄入会话。
-3. 当相邻会话跨越配置的时刻（默认 23:00）时触发 `auto_dream`。
+3. 若自定义应用配置启用了 `auto_dream`，在相邻会话跨越配置时刻（默认 23:00）时触发；插件预设保持关闭。
 4. 以 agentic（ReAct）模式回答每个问题。
 5. 通过 `answer_judge` 任务对答案做二元（yes/no）评判，并输出各类型准确率。

@ -55,7 +67,7 @@ python benchmark/longmemeval/run.py --eval_only               # 复用已有工
 | `dataset.workspace_root` | 条目工作区根目录（`benchmark/longmemeval/workspaces/longmemeval-s`）。 |
 | `evaluation.num_workers` | `0` = 自动（cpu-2），`1` = 串行，`>1` = 并行。 |
 | `evaluation.filter_future_sessions` | 仅摄入时间戳 ≤ `question_date` 的会话。 |
-| `reme.config` | 使用的 ReMe 配置（`lme.yaml`）。 |
+| `reme.config` | 使用的 ReMe 配置（`benchmark`）。 |
 | `reme.dream_trigger_hour` / `dream_scan_days` / `dream_max_units` | dream 触发行为。 |
 | `output.dir` | 结果目录（`benchmark/longmemeval/results`）。 |

--- a/benchmark/longmemeval/config.yaml
+++ b/benchmark/longmemeval/config.yaml
@ -10,7 +10,7 @@ dataset:
  workspace_root: "benchmark/longmemeval/workspaces/longmemeval-s"  # workspace root for item workspaces

 evaluation:
-  # LLM-as-judge uses the 'judge' as_llm component defined in lme.yaml
+  # LLM-as-judge uses the 'judge' as_llm component defined in benchmark.yaml
  # Model and credentials are configured there (reading from .env)
  # Judgment is always binary (yes/no) — defined in lme/llm_judge.yaml
  num_workers: 32            # 0 = auto (cpu_count - 2, min 1); 1 = sequential; >1 = parallel
@ -18,7 +18,7 @@ evaluation:
  compress_session: false   # true = compress session chunks in search_v2 (query-aware); false = no compression

 reme:
-  config: "lme.yaml" # reme config to use (in reme/config/)
+  config: "benchmark" # shared ReMe benchmark preset; runner enables the installed lme plugin
  # Dream trigger: when gap between consecutive sessions crosses this hour (23:00)
  dream_trigger_hour: 23
  # Dream scan_days for each trigger
--- a/benchmark/longmemeval/run.py
+++ b/benchmark/longmemeval/run.py
@ -28,7 +28,7 @@ import yaml
 from dotenv import load_dotenv

 # Load .env from project root
-_PROJECT_ROOT = Path(__file__).parent.parent.parent
+_PROJECT_ROOT = Path(__file__).resolve().parent.parent.parent
 load_dotenv(_PROJECT_ROOT / ".env")

 # Workspace root for evaluation items — read from config.yaml (dataset.workspace_root)
@ -150,6 +150,22 @@ def load_eval_config(config_path: str | None = None) -> dict:
    return yaml.safe_load(raw)


+def create_reme_app(config: str = "benchmark", **overrides):
+    """Create an app with the installed LongMemEval plugin explicitly enabled.
+
+    Plugin discovery remains environment-based; editable installation keeps local
+    plugin source changes visible to every multiprocessing worker.
+    """
+    from reme import Application
+    from reme.config import resolve_app_config
+
+    enabled_plugins = list(overrides.pop("plugins", ()) or ())
+    if "lme" not in enabled_plugins:
+        enabled_plugins.append("lme")
+    app_config = resolve_app_config(config=config, plugins=enabled_plugins, **overrides)
+    return Application(**app_config)
+
+
 # ---------------------------------------------------------------------------
 # Date utilities
 # ---------------------------------------------------------------------------
@ -257,8 +273,6 @@ async def evaluate_item(item: dict, eval_config: dict, item_index: int, eval_onl
            using the existing workspace. Useful for re-evaluating different query
            configurations without re-ingesting sessions.
    """
-    from reme import Application
-    from reme.config import resolve_app_config
    from reme.utils.evaluation_interface import track_agent_token_usage, track_job_counts

    reme_cfg = eval_config["reme"]
@ -325,7 +339,7 @@ async def evaluate_item(item: dict, eval_config: dict, item_index: int, eval_onl
                force_init=True,
            )

-    cfg = resolve_app_config(
+    app = create_reme_app(
        config=reme_cfg["config"],
        workspace_dir=workspace_dir,
        log_to_console=output_cfg.get("log_to_console", True),
@ -333,7 +347,6 @@ async def evaluate_item(item: dict, eval_config: dict, item_index: int, eval_onl
        enable_logo=False,
    )

-    app = Application(**cfg)
    await app.start()

    try:
--- a/benchmark/pibench/README.md
+++ b/benchmark/pibench/README.md
@ -289,7 +289,7 @@ removed"). They need the executed tool calls in the trace. The pipeline:
 | user_agent / judger models | `config/models/reme.yaml` |
 | Agent system prompt | `bridge_reme.py` `build_system_prompt()` |
 | Memory retrieval limit/threshold | `--search-limit/--search-min-score` on the bridge command in `run_persona.sh` |
-| ReMe internal parameters | **Do not modify ReMe source**; write a dedicated config modeled on `reme/config/beam.yaml` and override via `resolve_app_config(config=...)` (see bridge `_init_reme_app`) |
+| ReMe internal parameters | **Do not modify ReMe source**; extend the built-in `benchmark` config and override via `resolve_app_config(config=...)` (see bridge `_init_reme_app`) |
 | Turn timeout / tool iteration cap | `config/models/reme.yaml` `run.turn_timeout`, `model.max_tool_iterations` |

 ## 11. Troubleshooting
--- a/benchmark/pibench/README_ZH.md
+++ b/benchmark/pibench/README_ZH.md
@ -255,7 +255,7 @@ grep -h "overall_average_score\|overall_proactiveness" \
 | user_agent / judger 模型 | `config/models/reme.yaml` |
 | agent system prompt | `bridge_reme.py` `build_system_prompt()` |
 | 记忆检索条数/阈值 | `run_persona.sh` bridge 启动命令的 `--search-limit/--search-min-score` |
-| ReMe 内部参数 | **不要改 ReMe 源码**；仿照 `reme/config/beam.yaml` 写专有配置，经 `resolve_app_config(config=...)` 覆盖（见 bridge `_init_reme_app`） |
+| ReMe 内部参数 | **不要改 ReMe 源码**；继承内置 `benchmark` 配置，并经 `resolve_app_config(config=...)` 覆盖（见 bridge `_init_reme_app`） |
 | 轮超时/工具迭代上限 | `config/models/reme.yaml` `run.turn_timeout`、`model.max_tool_iterations` |

 ## 11. 故障排查
--- a/docs/en/plugin_management.md
+++ b/docs/en/plugin_management.md
@ -176,6 +176,16 @@ When the application uses an MCP service, service-enabled plugin Jobs appear as
 Custom application configs must provide the plugin's runtime dependencies, including an `agent_wrapper.default` and
 the `search` and `read` Jobs used by Auto Fin.

+## Benchmark application presets
+
+The [LME](../../plugins/lme/README.md) and [BEAM](../../plugins/beam/README.md) plugins
+register their backends and plugin-owned Jobs in `plugin.yaml`. ReMe's built-in `benchmark`
+preset provides the shared core Jobs and components without inheriting `default`, so default
+background and cron jobs are not included. Install the selected benchmark plugin, then use
+`config=benchmark` together with `plugins=["lme"]` or `plugins=["beam"]`. The repository's
+benchmark runners enable the corresponding installed plugin automatically; editable installation
+keeps local plugin changes visible. Dataset runners remain under `benchmark/`.
+
 ## Uninstall a plugin

 Use the plugin entry-point name, not necessarily the distribution name:
--- a/docs/zh/plugin_management.md
+++ b/docs/zh/plugin_management.md
@ -169,6 +169,15 @@ curl -s http://127.0.0.1:2333/auto_fin \

 自定义应用配置需要提供插件的运行依赖，包括 `agent_wrapper.default`，以及 Auto Fin 使用的 `search` 和 `read` Jobs。

+## Benchmark 应用配置
+
+[LME](../../plugins/lme/README_ZH.md) 和 [BEAM](../../plugins/beam/README_ZH.md) 插件通过
+`plugin.yaml` 注册 backend 和插件拥有的 Job。ReMe 内置的 `benchmark` 配置提供公共核心 Job 和
+Component，并且不继承 `default`，因此不包含默认后台和定时任务。先安装所需的 benchmark 插件，
+再使用 `config=benchmark`，同时指定 `plugins=["lme"]` 或 `plugins=["beam"]`。仓库内的 benchmark
+runner 会自动启用对应的已安装插件；editable 安装可让本地源码修改直接生效。数据集 runner 仍位于
+`benchmark/`。
+
 ## 卸载插件

 这里使用插件 entry-point 名称，它不一定等于 distribution 名称：
--- a/plugins/beam/LICENSE
+++ b/plugins/beam/LICENSE
@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2025 Alibaba Group
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
--- a/plugins/beam/README.md
+++ b/plugins/beam/README.md
@ -0,0 +1,38 @@
+# BEAM plugin
+
+[中文说明](./README_ZH.md)
+
+This plugin owns the BEAM memory, agentic-answer and judge Steps, their prompts,
+and their Job defaults in `plugin.yaml`. ReMe's built-in `benchmark.yaml` owns the
+shared evaluation Jobs and components. Dataset handling, the runner and results
+remain in [`benchmark/beam`](../../benchmark/beam/README.md).
+
+From the repository root, install ReMe and this plugin in editable mode before running the benchmark:
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/beam --editable
+reme plugins validate beam
+python benchmark/beam/run.py
+```
+
+Editable installation registers the `beam` entry point while keeping source changes immediately
+visible. The runner selects the built-in `benchmark` preset and explicitly enables `beam` for
+each Application. Installing the plugin makes it discoverable but does not enable it globally.
+
+`plugin.yaml` registers backends and contributes the plugin-owned `auto_memory`,
+`agentic_answer` and `answer_judge` Job defaults. Start the installed plugin with
+`reme start config=benchmark plugins='["beam"]'`. The shared preset does not inherit
+`default`: only declared Jobs run, indexing is manual, and neither scheduled dream
+nor the optional `auto_dream` Job is enabled.
+The existing `auto_memory`, `agentic_answer`, `answer_judge`, `bench` and `judge`
+names and model environment variables are unchanged. Explicit application/CLI overrides
+still take precedence. Installing this plugin does not start an evaluation.
+
+The shared answer base class lives in `reme.steps.benchmark.base_agentic_answer`.
+The old core-owned `reme.steps.benchmark.beam` Python import path is removed.
+Custom Python callers should import memory, search and answer Steps from `reme_beam`, and the
+judge Step from `judge_beam`. After uninstalling,
+Applications and CLI services must omit the plugin until it is installed again.
+Uninstallation never removes datasets, workspaces or results.
+Restart an existing service after changing plugins.
--- a/plugins/beam/README_ZH.md
+++ b/plugins/beam/README_ZH.md
@ -0,0 +1,32 @@
+# BEAM 插件
+
+[English](./README.md)
+
+插件包含 BEAM 的记忆、回答、评分 Step、提示词，以及 `plugin.yaml` 中对应的 Job 默认配置。
+ReMe 内置的 `benchmark.yaml` 负责公共评测 Job 和 Component；数据集处理、runner 和结果仍留在
+[`benchmark/beam`](../../benchmark/beam/README_ZH.md)。
+
+在仓库根目录以 editable 模式安装 ReMe 和本插件，再运行评测：
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/beam --editable
+reme plugins validate beam
+python benchmark/beam/run.py
+```
+
+editable 安装会注册 `beam` entry point，并让源码修改立即生效。runner 选择内置 `benchmark`
+配置，并为每个 Application 显式启用 `beam`。安装只让插件可被发现，不会在所有应用中全局启用。
+
+`plugin.yaml` 注册 backend，并通过 `application_defaults` 提供插件拥有的 `auto_memory`、
+`agentic_answer` 和 `answer_judge` Job。安装后使用
+`reme start config=benchmark plugins='["beam"]'`。公共评测配置不继承 `default`，只运行声明的 Job：
+索引手动更新，dream 定时任务和可选的 `auto_dream`
+均保持关闭。原有 `auto_memory`、`agentic_answer`、`answer_judge`、`bench`、`judge` 名称及模型环境变量
+保持不变，显式应用参数和 CLI 覆盖仍优先。安装或启用插件不会自动开始评测。
+
+共享回答基类位于 `reme.steps.benchmark.base_agentic_answer`。
+原 `reme.steps.benchmark.beam` Python 导入路径已移除。自定义 Python 调用应从 `reme_beam`
+导入记忆、搜索和回答 Step，并从 `judge_beam` 导入评判 Step。
+卸载插件后，Application 和 CLI 服务必须移除插件选择，直到再次安装。
+卸载不会删除数据集、工作区或结果。修改插件后需重启已有服务。
--- a/plugins/beam/pyproject.toml
+++ b/plugins/beam/pyproject.toml
@ -0,0 +1,28 @@
+[project]
+name = "reme-beam"
+version = "0.1.0"
+description = "BEAM benchmark plugin for ReMe."
+readme = "README.md"
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+requires-python = ">=3.11"
+dependencies = [
+    "reme-ai[as]>=0.4.1.11",
+    "json-repair",
+    "numpy>=2.2.6",
+]
+
+[project.entry-points."reme.plugins"]
+beam = "reme_beam"
+
+[tool.setuptools.packages.find]
+where = ["src"]
+include = ["reme_beam*", "judge_beam*"]
+
+[tool.setuptools.package-data]
+reme_beam = ["plugin.yaml", "*.yaml"]
+judge_beam = ["*.yaml"]
+
+[build-system]
+requires = ["setuptools>=77", "wheel"]
+build-backend = "setuptools.build_meta"
--- a/plugins/beam/src/judge_beam/init.py
+++ b/plugins/beam/src/judge_beam/init.py
@ -0,0 +1,5 @@
+"""BEAM benchmark judge backend."""
+
+from .llm_judge import BeamRubricJudgeStep
+
+__all__ = ["BeamRubricJudgeStep"]
--- a/plugins/beam/src/judge_beam/llm_judge.py
+++ b/plugins/beam/src/judge_beam/llm_judge.py
@ -24,10 +24,9 @@ from typing import List, Tuple
 import numpy as np
 from json_repair import repair_json

-from ...base_step import BaseStep, Ref
-from ....components import R
-from ....components.as_embedding import BaseAsEmbedding
-from ....enumeration import ComponentEnum
+from reme.steps.base_step import BaseStep, Ref
+from reme.components.as_embedding import BaseAsEmbedding
+from reme.enumeration import ComponentEnum


 # ---------------------------------------------------------------------------
@ -231,7 +230,6 @@ def _event_ordering_score(
    }


-@R.register("beam_rubric_judge_step")
 class BeamRubricJudgeStep(BaseStep):
    """Judge an LLM response against a list of rubric criteria.

--- a/plugins/beam/src/judge_beam/llm_judge.yaml
+++ b/plugins/beam/src/judge_beam/llm_judge.yaml
--- a/plugins/beam/src/reme_beam/init.py
+++ b/plugins/beam/src/reme_beam/init.py
@ -1,11 +1,11 @@
-"""BEAM benchmark steps."""
+"""BEAM benchmark backends and application configuration for ReMe."""

 from .agentic_answer import BeamAgenticAnswerStep
-from .llm_judge import BeamRubricJudgeStep
 from .auto_memory import BeamAutoMemoryStep
+from .search_v2 import SearchV2Step

 __all__ = [
    "BeamAgenticAnswerStep",
-    "BeamRubricJudgeStep",
    "BeamAutoMemoryStep",
+    "SearchV2Step",
 ]
--- a/plugins/beam/src/reme_beam/agentic_answer.py
+++ b/plugins/beam/src/reme_beam/agentic_answer.py
@ -1,10 +1,8 @@
 """BEAM agentic answer step – ReAct agent that answers questions using the search tool."""

-from ....components import R
-from ..base import BaseAgenticAnswerStep
+from reme.steps.benchmark import BaseAgenticAnswerStep


-@R.register("beam_agentic_answer_step")
 class BeamAgenticAnswerStep(BaseAgenticAnswerStep):
    """Answer a BEAM probing question via ReAct agent with access to the search tool.

--- a/plugins/beam/src/reme_beam/agentic_answer.yaml
+++ b/plugins/beam/src/reme_beam/agentic_answer.yaml
--- a/plugins/beam/src/reme_beam/auto_memory.py
+++ b/plugins/beam/src/reme_beam/auto_memory.py
@ -4,9 +4,8 @@ from datetime import datetime, timedelta

 from agentscope.message import Msg

-from ...evolve.auto_memory import AutoMemoryStep, _normalize_msg_timestamp
-from ...file_io import validate_session_id
-from ....components import R
+from reme.steps.evolve.auto_memory import AutoMemoryStep, _normalize_msg_timestamp
+from reme.steps.file_io import validate_session_id

 # Runtime-context key carrying the 0-based line offset of the current segment
 # inside the full session file (segmented ingestion of long sessions).
@ -168,7 +167,6 @@ def _interpolate_timestamps(items: list[dict]) -> list[dict]:
    return result


-@R.register("beam_auto_memory_step")
 class BeamAutoMemoryStep(AutoMemoryStep):
    """AutoMemoryStep variant that interpolates timestamps for BEAM sessions.

--- a/plugins/beam/src/reme_beam/auto_memory.yaml
+++ b/plugins/beam/src/reme_beam/auto_memory.yaml
--- a/plugins/beam/src/reme_beam/plugin.yaml
+++ b/plugins/beam/src/reme_beam/plugin.yaml
@ -0,0 +1,120 @@
+backends:
+  beam_auto_memory_step: reme_beam.auto_memory:BeamAutoMemoryStep
+  beam_agentic_answer_step: reme_beam.agentic_answer:BeamAgenticAnswerStep
+  beam_rubric_judge_step: judge_beam.llm_judge:BeamRubricJudgeStep
+  beam_search_v2_step: reme_beam.search_v2:SearchV2Step
+
+application_defaults:
+  jobs:
+    search:
+      backend: base
+      description: "Hybrid workspace search (vector + BM25, RRF-fused) with deduplication."
+      parameters:
+        type: object
+        properties:
+          query:
+            type: string
+            description: "search query"
+          start_date:
+            type: string
+            description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
+          end_date:
+            type: string
+            description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
+        required:
+          - query
+      steps:
+        - backend: beam_search_v2_step
+          vector_weight: 0.7
+          candidate_multiplier: 5.0
+          expand_links: false
+          max_links_per_direction: 10
+
+    agentic_answer:
+      backend: base
+      description: "BEAM agentic answer job (ReAct agent with search tool)"
+      watch_dirs: []
+      watch_suffixes: []
+      parameters:
+        type: object
+        properties:
+          query:
+            type: string
+            description: "The query to ask"
+          query_time:
+            type: string
+            description: "ISO timestamp representing the query time"
+            default: ""
+        required:
+          - query
+      steps:
+        - backend: beam_agentic_answer_step
+          agent_wrapper: bench
+
+    answer_judge:
+      backend: base
+      description: "BEAM rubric-based LLM-as-Judge: evaluate response against rubric criteria"
+      watch_dirs: []
+      watch_suffixes: []
+      parameters:
+        type: object
+        properties:
+          llm_response:
+            type: string
+            description: "The model's response to evaluate"
+          rubric:
+            type: array
+            description: "List of rubric criteria to check"
+            items:
+              type: string
+          probing_question:
+            type: string
+            description: "The original probing question"
+            default: ""
+          question_type:
+            type: string
+            description: "BEAM question type (e.g. event_ordering)"
+            default: ""
+        required:
+          - llm_response
+          - rubric
+      steps:
+        - backend: beam_rubric_judge_step
+          agent_wrapper: judge
+
+    auto_memory:
+      backend: base
+      description: "Auto-memory: record conversation facts into a daily note"
+      parameters:
+        type: object
+        properties:
+          messages:
+            type: array
+            description: "messages"
+            items:
+              type: object
+          session_id:
+            type: string
+            description: "source conversation session identifier"
+            default: ""
+          memory_hint:
+            type: string
+            description: "optional hint"
+          date:
+            type: string
+            description: "YYYY-MM-DD daily note date; empty = infer from message timestamps or today"
+            default: ""
+        required:
+          - messages
+      steps:
+        - backend: beam_auto_memory_step
+          # Long sessions are split into turn-aligned segments and fed to the
+          # agent incrementally; <= 0 disables splitting.
+          max_segment_words: 10000
+
+  components:
+    as_llm:
+      default:
+        max_retries: 5
+      judge:
+        retry_delay: 5.0
--- a/plugins/beam/src/reme_beam/search_v2.py
+++ b/plugins/beam/src/reme_beam/search_v2.py
@ -12,14 +12,18 @@ import datetime
 import os
 from typing import Final

-from ._dedup import _ToolContextDedupMixin
-from ._source_format import ALL_RETURNED_MESSAGE, NO_RESULTS_MESSAGE, is_session_path, join_chunk_entries
-from ._source_format import merge_session_chunk_intervals, render_chunk_entries
-from ..base_step import BaseStep
-from ..file_io import extract_daily_date
-from ...components import R
-from ...schema import FileChunk
-from ...utils import expand_links
+from reme.schema import FileChunk
+from reme.steps.base_step import BaseStep
+from reme.steps.file_io import extract_daily_date
+from reme.steps.index._dedup import _ToolContextDedupMixin
+from reme.steps.index._source_format import (
+    ALL_RETURNED_MESSAGE,
+    NO_RESULTS_MESSAGE,
+    is_session_path,
+    join_chunk_entries,
+)
+from reme.steps.index._source_format import merge_session_chunk_intervals, render_chunk_entries
+from reme.utils import expand_links

 _RRF_K: Final = 60
 _MAX_CANDIDATES: Final = 200
@ -37,7 +41,6 @@ def _default_limit() -> int:
        return _DEFAULT_LIMIT


-@R.register("search_v2_step")
 class SearchV2Step(_ToolContextDedupMixin, BaseStep):
    """Hybrid search: run vector + keyword in parallel, fuse via RRF, filter, truncate."""

--- a/plugins/lme/LICENSE
+++ b/plugins/lme/LICENSE
@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2025 Alibaba Group
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
--- a/plugins/lme/README.md
+++ b/plugins/lme/README.md
@ -0,0 +1,38 @@
+# LongMemEval plugin
+
+[中文说明](./README_ZH.md)
+
+This plugin owns the LongMemEval memory, agentic-answer and judge Steps, their prompts,
+and their Job defaults in `plugin.yaml`. ReMe's built-in `benchmark.yaml` owns the
+shared evaluation Jobs and components. Dataset handling, the runner and results remain
+in [`benchmark/longmemeval`](../../benchmark/longmemeval/README.md).
+
+From the repository root, install ReMe and this plugin in editable mode before running the benchmark:
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/lme --editable
+reme plugins validate lme
+python benchmark/longmemeval/run.py
+```
+
+Editable installation registers the `lme` entry point while keeping source changes immediately
+visible. The runner selects the built-in `benchmark` preset and explicitly enables `lme` for
+each Application. Installing the plugin makes it discoverable but does not enable it globally.
+
+`plugin.yaml` registers backends and contributes the plugin-owned `auto_memory`,
+`agentic_answer` and `answer_judge` Job defaults. Start the installed plugin with
+`reme start config=benchmark plugins='["lme"]'`. The shared preset does not inherit
+`default`: only declared Jobs run, indexing is manual, and neither scheduled dream
+nor the optional `auto_dream` Job is enabled.
+The existing `auto_memory`, `agentic_answer`, `answer_judge`, `bench` and `judge`
+names and model environment variables are unchanged. Explicit application/CLI overrides
+still take precedence. Installing this plugin does not start an evaluation.
+
+The shared answer base class lives in `reme.steps.benchmark.base_agentic_answer`.
+The old core-owned `reme.steps.benchmark.lme` Python import path is removed.
+Custom Python callers should import memory, search and answer Steps from `reme_lme`, and the
+judge Step from `judge_lme`. After uninstalling,
+Applications and CLI services must omit the plugin until it is installed again.
+Uninstallation never removes datasets, workspaces or results.
+Restart an existing service after changing plugins.
--- a/plugins/lme/README_ZH.md
+++ b/plugins/lme/README_ZH.md
@ -0,0 +1,32 @@
+# LongMemEval 插件
+
+[English](./README.md)
+
+插件包含 LongMemEval 的记忆、回答、评分 Step、提示词，以及 `plugin.yaml` 中对应的 Job 默认配置。
+ReMe 内置的 `benchmark.yaml` 负责公共评测 Job 和 Component；数据集处理、runner 和结果仍留在
+[`benchmark/longmemeval`](../../benchmark/longmemeval/README_ZH.md)。
+
+在仓库根目录以 editable 模式安装 ReMe 和本插件，再运行评测：
+
+```bash
+python -m pip install -e ".[as]"
+reme plugins install ./plugins/lme --editable
+reme plugins validate lme
+python benchmark/longmemeval/run.py
+```
+
+editable 安装会注册 `lme` entry point，并让源码修改立即生效。runner 选择内置 `benchmark`
+配置，并为每个 Application 显式启用 `lme`。安装只让插件可被发现，不会在所有应用中全局启用。
+
+`plugin.yaml` 注册 backend，并通过 `application_defaults` 提供插件拥有的 `auto_memory`、
+`agentic_answer` 和 `answer_judge` Job。安装后使用
+`reme start config=benchmark plugins='["lme"]'`。公共评测配置不继承 `default`，只运行声明的 Job：
+索引手动更新，dream 定时任务和可选的 `auto_dream`
+均保持关闭。原有 `auto_memory`、`agentic_answer`、`answer_judge`、`bench`、`judge` 名称及模型环境变量
+保持不变，显式应用参数和 CLI 覆盖仍优先。安装或启用插件不会自动开始评测。
+
+共享回答基类位于 `reme.steps.benchmark.base_agentic_answer`。
+原 `reme.steps.benchmark.lme` Python 导入路径已移除。自定义 Python 调用应从 `reme_lme`
+导入记忆、搜索和回答 Step，并从 `judge_lme` 导入评判 Step。
+卸载插件后，Application 和 CLI 服务必须移除插件选择，直到再次安装。
+卸载不会删除数据集、工作区或结果。修改插件后需重启已有服务。
--- a/plugins/lme/pyproject.toml
+++ b/plugins/lme/pyproject.toml
@ -0,0 +1,26 @@
+[project]
+name = "reme-lme"
+version = "0.1.0"
+description = "LongMemEval benchmark plugin for ReMe."
+readme = "README.md"
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+requires-python = ">=3.11"
+dependencies = [
+    "reme-ai[as]>=0.4.1.11",
+]
+
+[project.entry-points."reme.plugins"]
+lme = "reme_lme"
+
+[tool.setuptools.packages.find]
+where = ["src"]
+include = ["reme_lme*", "judge_lme*"]
+
+[tool.setuptools.package-data]
+reme_lme = ["plugin.yaml", "*.yaml"]
+judge_lme = ["*.yaml"]
+
+[build-system]
+requires = ["setuptools>=77", "wheel"]
+build-backend = "setuptools.build_meta"
--- a/plugins/lme/src/judge_lme/init.py
+++ b/plugins/lme/src/judge_lme/init.py
@ -0,0 +1,5 @@
+"""LongMemEval benchmark judge backend."""
+
+from .llm_judge import LmeAnswerJudgeStep
+
+__all__ = ["LmeAnswerJudgeStep"]
--- a/plugins/lme/src/judge_lme/llm_judge.py
+++ b/plugins/lme/src/judge_lme/llm_judge.py
@ -2,11 +2,9 @@

 import re

-from ...base_step import BaseStep
-from ....components import R
+from reme.steps.base_step import BaseStep


-@R.register("lme_answer_judge_step")
 class LmeAnswerJudgeStep(BaseStep):
    """Evaluate whether an agent answer is correct against a golden answer."""

--- a/plugins/lme/src/judge_lme/llm_judge.yaml
+++ b/plugins/lme/src/judge_lme/llm_judge.yaml
--- a/plugins/lme/src/reme_lme/init.py
+++ b/plugins/lme/src/reme_lme/init.py
@ -1,11 +1,11 @@
-"""LongMemEval benchmark steps."""
+"""LongMemEval benchmark backends and application configuration for ReMe."""

 from .agentic_answer import LmeAgenticAnswerStep
-from .llm_judge import LmeAnswerJudgeStep
 from .auto_memory import LmeAutoMemoryStep
+from .search_v2 import SearchV2Step

 __all__ = [
    "LmeAgenticAnswerStep",
-    "LmeAnswerJudgeStep",
    "LmeAutoMemoryStep",
+    "SearchV2Step",
 ]
--- a/plugins/lme/src/reme_lme/agentic_answer.py
+++ b/plugins/lme/src/reme_lme/agentic_answer.py
@ -1,10 +1,8 @@
 """LongMemEval agentic answer step – ReAct agent that answers questions using the search tool."""

-from ....components import R
-from ..base import BaseAgenticAnswerStep
+from reme.steps.benchmark import BaseAgenticAnswerStep


-@R.register("lme_agentic_answer_step")
 class LmeAgenticAnswerStep(BaseAgenticAnswerStep):
    """Answer a LongMemEval query via ReAct agent with access to the search tool.

@ -12,7 +10,7 @@ class LmeAgenticAnswerStep(BaseAgenticAnswerStep):
    ``search`` job tool to retrieve relevant memory chunks before generating
    a final answer.

-    Session-transcript compression in ``search_v2_step`` is controlled by the
+    Session-transcript compression in the plugin's search Step is controlled by the
    ``compress_session`` flag in the runtime context (set by the benchmark
    runner from ``evaluation.compress_session``); it is off by default.
    """
--- a/plugins/lme/src/reme_lme/agentic_answer.yaml
+++ b/plugins/lme/src/reme_lme/agentic_answer.yaml
--- a/plugins/lme/src/reme_lme/auto_memory.py
+++ b/plugins/lme/src/reme_lme/auto_memory.py
@ -4,8 +4,7 @@ from datetime import datetime, timedelta

 from agentscope.message import Msg

-from ...evolve.auto_memory import AutoMemoryStep, _normalize_msg_timestamp
-from ....components import R
+from reme.steps.evolve.auto_memory import AutoMemoryStep, _normalize_msg_timestamp


 def _parse_iso_seconds(value: str) -> datetime | None:
@ -115,7 +114,6 @@ def _interpolate_timestamps(items: list[dict]) -> list[dict]:
    return result


-@R.register("lme_auto_memory_step")
 class LmeAutoMemoryStep(AutoMemoryStep):
    """AutoMemoryStep variant that interpolates timestamps for LongMemEval sessions.

--- a/plugins/lme/src/reme_lme/auto_memory.yaml
+++ b/plugins/lme/src/reme_lme/auto_memory.yaml
--- a/plugins/lme/src/reme_lme/plugin.yaml
+++ b/plugins/lme/src/reme_lme/plugin.yaml
@ -0,0 +1,113 @@
+backends:
+  lme_auto_memory_step: reme_lme.auto_memory:LmeAutoMemoryStep
+  lme_agentic_answer_step: reme_lme.agentic_answer:LmeAgenticAnswerStep
+  lme_answer_judge_step: judge_lme.llm_judge:LmeAnswerJudgeStep
+  lme_search_v2_step: reme_lme.search_v2:SearchV2Step
+
+application_defaults:
+  jobs:
+    search:
+      backend: base
+      description: "Hybrid workspace search (vector + BM25, RRF-fused) with deduplication."
+      parameters:
+        type: object
+        properties:
+          query:
+            type: string
+            description: "search query"
+          start_date:
+            type: string
+            description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
+          end_date:
+            type: string
+            description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
+        required:
+          - query
+      steps:
+        - backend: lme_search_v2_step
+          vector_weight: 0.7
+          candidate_multiplier: 5.0
+          expand_links: false
+          max_links_per_direction: 10
+
+    agentic_answer:
+      backend: base
+      description: "LongMemEval agentic answer job (ReAct agent with search tool)"
+      watch_dirs: []
+      watch_suffixes: []
+      parameters:
+        type: object
+        properties:
+          query:
+            type: string
+            description: "The query to ask"
+          query_time:
+            type: string
+            description: "ISO timestamp representing the query time"
+            default: ""
+        required:
+          - query
+      steps:
+        - backend: lme_agentic_answer_step
+          agent_wrapper: bench
+
+    answer_judge:
+      backend: base
+      description: "LLM-as-Judge: evaluate agent answer against golden answer"
+      watch_dirs: []
+      watch_suffixes: []
+      parameters:
+        type: object
+        properties:
+          query:
+            type: string
+            description: "The question being asked"
+          agent_answer:
+            type: string
+            description: "The model's answer to evaluate"
+          golden_answer:
+            type: string
+            description: "The correct/golden answer"
+          question_type:
+            type: string
+            description: "The question type for prompt selection"
+            default: ""
+        required:
+          - query
+          - agent_answer
+          - golden_answer
+      steps:
+        - backend: lme_answer_judge_step
+          agent_wrapper: judge
+
+    auto_memory:
+      backend: base
+      description: "Auto-memory: record conversation facts into a daily note"
+      parameters:
+        type: object
+        properties:
+          messages:
+            type: array
+            description: "messages"
+            items:
+              type: object
+          session_id:
+            type: string
+            description: "source conversation session identifier"
+            default: ""
+          memory_hint:
+            type: string
+            description: "optional hint"
+          date:
+            type: string
+            description: "YYYY-MM-DD daily note date; empty = infer from message timestamps or today"
+            default: ""
+        required:
+          - messages
+      steps:
+        - backend: lme_auto_memory_step
+
+  components:
+    as_llm:
+      default:
+        max_retries: 3
--- a/plugins/lme/src/reme_lme/search_v2.py
+++ b/plugins/lme/src/reme_lme/search_v2.py
@ -0,0 +1,333 @@
+"""Hybrid search (v2) over file_store using RRF fusion of vector + keyword results.
+
+This is the local fork of the upstream search step. It uses
+:class:`_ToolContextDedupMixin` for subset-aware interval-merging dedup and
+:func:`render_chunk_entries` / :func:`join_chunk_entries` for session-aware
+chunk formatting with
+:data:`ALL_RETURNED_MESSAGE` / :data:`NO_RESULTS_MESSAGE` notices.
+"""
+
+import asyncio
+import datetime
+import os
+from typing import Final
+
+from reme.schema import FileChunk
+from reme.steps.base_step import BaseStep
+from reme.steps.file_io import extract_daily_date
+from reme.steps.index._dedup import _ToolContextDedupMixin
+from reme.steps.index._source_format import (
+    ALL_RETURNED_MESSAGE,
+    NO_RESULTS_MESSAGE,
+    is_session_path,
+    join_chunk_entries,
+)
+from reme.steps.index._source_format import merge_session_chunk_intervals, render_chunk_entries
+from reme.utils import expand_links
+
+_RRF_K: Final = 60
+_MAX_CANDIDATES: Final = 200
+_DEFAULT_LIMIT_ENV: Final = "REME_SEARCH_LIMIT"
+_DEFAULT_LIMIT: Final = 5
+
+
+def _default_limit() -> int:
+    value = os.getenv(_DEFAULT_LIMIT_ENV)
+    if value is None:
+        return _DEFAULT_LIMIT
+    try:
+        return int(value)
+    except ValueError:
+        return _DEFAULT_LIMIT
+
+
+class SearchV2Step(_ToolContextDedupMixin, BaseStep):
+    """Hybrid search: run vector + keyword in parallel, fuse via RRF, filter, truncate."""
+
+    def __init__(
+        self,
+        *args,
+        seen_ttl_hours: float = 24,
+        **kwargs,
+    ):
+        super().__init__(*args, **kwargs)
+        self.seen_ttl_hours = seen_ttl_hours
+
+    @staticmethod
+    def _rrf_merge(
+        vector: list[FileChunk],
+        keyword: list[FileChunk],
+        vector_weight: float,
+    ) -> list[FileChunk]:
+        """Fuse two ranked lists with Reciprocal Rank Fusion, keyed by chunk.id."""
+        text_weight = 1.0 - vector_weight
+        merged: dict[str, FileChunk] = {}
+
+        for rank, chunk in enumerate(vector, start=1):
+            contrib = vector_weight / (_RRF_K + rank)
+            c = chunk.model_copy(deep=False)
+            c.scores = {**chunk.scores, "vector": chunk.scores.get("vector", chunk.score), "score": contrib}
+            merged[c.id] = c
+
+        for rank, chunk in enumerate(keyword, start=1):
+            contrib = text_weight / (_RRF_K + rank)
+            existing = merged.get(chunk.id)
+            if existing is not None:
+                existing.scores = {
+                    **existing.scores,
+                    "keyword": chunk.scores.get("keyword", chunk.score),
+                    "score": existing.scores["score"] + contrib,
+                }
+            else:
+                c = chunk.model_copy(deep=False)
+                c.scores = {**chunk.scores, "keyword": chunk.scores.get("keyword", chunk.score), "score": contrib}
+                merged[c.id] = c
+
+        results = list(merged.values())
+        results.sort(key=lambda r: r.score, reverse=True)
+        return results
+
+    @staticmethod
+    def _format_scores(scores: dict[str, float], hybrid: bool) -> str:
+        """Format scores for the answer line: always show fused; show per-branch when hybrid."""
+        parts = [f"score={scores.get('score', 0.0):.4f}"]
+        if hybrid:
+            for k in ("vector", "keyword"):
+                v = scores.get(k)
+                parts.append(f"{k}={v:.4f}" if v is not None else f"{k}=-")
+        return " ".join(parts)
+
+    async def execute(self):
+        assert self.context is not None
+        query: str = (self.context.get("query", "") or "").strip()
+        limit: int = int(self.context.get("limit") or _default_limit())
+        min_score: float = float(self.context.get("min_score") or 0.0)
+        # vector_weight: prefer agent-supplied context value; fallback to YAML kwargs / default 0.7.
+        # Convertible numeric inputs are clipped to [0.0, 1.0]; non-numeric inputs are silently ignored.
+        raw_vw = self.context.get("vector_weight")
+        vector_weight: float | None = None
+        if raw_vw is not None:
+            try:
+                vector_weight = float(raw_vw)
+            except (TypeError, ValueError):
+                self.logger.warning(
+                    f"[{self.name}] non-numeric vector_weight={raw_vw!r}; ignoring and using default 0.7",
+                )
+                vector_weight = None
+        if vector_weight is None:
+            vector_weight = float(self.kwargs.get("vector_weight", 0.7))
+        vector_weight = max(0.0, min(1.0, vector_weight))
+        candidate_multiplier: float = float(self.kwargs.get("candidate_multiplier", 5.0))
+        expand_links_enabled: bool = bool(self.kwargs.get("expand_links", True))
+        max_links_per_direction: int = int(self.kwargs.get("max_links_per_direction", 10))
+        tool_context_id: str = (self.context.get("tool_context_id", "") or "").strip()
+        # Injected value takes precedence over YAML kwargs; check existence
+        # (not truthiness) so an explicit False can disable a YAML-true flag.
+        _strict_date_filter = self.context.get("strict_date_filter")
+        if _strict_date_filter is None:
+            _strict_date_filter = self.kwargs.get("strict_date_filter", False)
+        strict_date_filter: bool = bool(_strict_date_filter)
+
+        if not query:
+            self.context.response.success = False
+            self.context.response.answer = "Error: query cannot be empty"
+            return self.context.response
+        assert limit > 0, f"limit must be positive, got {limit}"
+
+        candidates = min(_MAX_CANDIDATES, max(1, int(limit * candidate_multiplier)))
+        search_filter: dict = dict(self.context.get("search_filter", {}) or {})
+
+        # Promote top-level date parameters into search_filter for file_store.
+        for date_key in ("start_date", "end_date"):
+            value = self.context.get(date_key)
+            if value and date_key not in search_filter:
+                search_filter[date_key] = value
+
+        # Validate and normalize date filters before they reach file_store.
+        # _matches_search_filter does lexicographic string comparison against
+        # path_date (always a canonical YYYY-MM-DD), so raw caller values like
+        # "2026-2-28" or "abc" would produce silently wrong results.
+        for date_key in ("start_date", "end_date"):
+            raw = search_filter.get(date_key)
+            if raw is None:
+                continue
+            normalized = extract_daily_date(raw)
+            if normalized is None:
+                # Fallback: accept non-zero-padded dates like "2024-1-5".
+                try:
+                    normalized = (
+                        datetime.datetime.strptime(
+                            str(raw).strip(),
+                            "%Y-%m-%d",
+                        )
+                        .date()
+                        .isoformat()
+                    )
+                except ValueError:
+                    self.logger.warning(
+                        f"Ignoring invalid {date_key}={raw!r}; " f"expected a valid YYYY-MM-DD date.",
+                    )
+                    del search_filter[date_key]
+                    continue
+            search_filter[date_key] = normalized
+
+        if strict_date_filter:
+            search_filter["strict_date_filter"] = True
+
+        vector_results, keyword_results = await asyncio.gather(
+            self.file_store.vector_search(query, candidates, search_filter),
+            self.file_store.keyword_search(query, candidates, search_filter),
+        )
+
+        self.logger.info(
+            f"[{self.name}] query={query!r} candidates={candidates} "
+            f"vector_hits={len(vector_results)} keyword_hits={len(keyword_results)}",
+        )
+
+        hybrid = bool(vector_results) and bool(keyword_results)
+        if not vector_results and not keyword_results:
+            fused: list[FileChunk] = []
+        elif not keyword_results:
+            fused = vector_results
+        elif not vector_results:
+            fused = keyword_results
+        else:
+            fused = self._rrf_merge(vector_results, keyword_results, vector_weight)
+
+        if min_score > 0.0:
+            fused = [c for c in fused if c.score >= min_score]
+
+        pre_dedup_count = 0
+        dedup: dict | None = None
+        if tool_context_id:
+            pre_dedup_count = len(fused)
+            fused, dedup = self._dedupe_tool_context(
+                fused,
+                tool_context_id,
+                limit,
+                clock=self.kwargs.get("clock"),
+                ttl_override=self.kwargs.get("tool_context_chunk_ttl_seconds"),
+            )
+        else:
+            fused = fused[:limit]
+
+        unique_paths = list(dict.fromkeys(c.path for c in fused))
+        link_expansion: dict[str, dict] = (
+            await expand_links(self.file_store, unique_paths, max_links_per_direction) if expand_links_enabled else {}
+        )
+
+        session_dir = self.config_value("session_dir")
+        entries = render_chunk_entries(
+            merge_session_chunk_intervals(fused, session_dir),
+            session_dir,
+            score_fn=lambda c: self._format_scores(c.scores, hybrid),
+            link_expansion=link_expansion,
+        )
+        if self._session_compress_enabled():
+            await self._compress_session_entries(entries, query, session_dir)
+        self.context.response.answer = join_chunk_entries(entries)
+        if not fused:
+            self.context.response.answer = ALL_RETURNED_MESSAGE if pre_dedup_count > 0 else NO_RESULTS_MESSAGE
+        self.context.response.metadata["results"] = [
+            c.model_dump(exclude_none=True, exclude={"embedding"}) for c in fused
+        ]
+        self.context.response.metadata["link_expansion"] = link_expansion
+        self.context.response.metadata["counts"] = {
+            "vector": len(vector_results),
+            "keyword": len(keyword_results),
+            "returned": len(fused),
+            "hybrid": hybrid,
+        }
+        if dedup is not None:
+            self.context.response.metadata["dedup"] = dedup
+        return self.context.response
+
+    def _session_compress_enabled(self) -> bool:
+        """True when the injected ``_search._compress.session`` flag is truthy."""
+        assert self.context is not None
+        search_cfg: dict = self.context.get("_search") or {}
+        value = (search_cfg.get("_compress") or {}).get("session")
+        return value is True or str(value).strip().lower() == "true"
+
+    async def _compress_session_entries(self, entries: list[dict[str, str]], query: str, session_dir: str) -> None:
+        """Compress session-transcript entry bodies in place via the ``compressor`` job.
+
+        Only entries whose ``path`` points at a raw session transcript are
+        compressed; other entries and all non-``body`` fields stay untouched.
+        When ``_search.type`` is ``query-independent`` the compressor runs
+        without queries (generic compression); otherwise (``query-aware``,
+        the default) it receives the injected ``_search.queries`` plus the
+        current search query.
+
+        The compressor receives the already-rendered body (one message per
+        line) stripped. Its output is adopted whenever the compressor
+        succeeded and the result is not longer than the input; adopted bodies
+        get a leading ``compressed session chunk:`` marker so downstream
+        consumers can tell them from verbatim transcripts.
+
+        Degrades gracefully when the ``compressor`` job is missing from the
+        active config: the whole method becomes a no-op and a warning is
+        logged, so search behaves as if compression were disabled. This
+        avoids a hard ``Job compressor not found`` failure when a benchmark
+        config forgets to define the compressor job/component.
+
+        Per-entry exceptions raised by the compressor job (e.g. a temporary
+        LLM outage) are caught inside ``compress`` so they never propagate
+        through ``asyncio.gather``: the failing entry keeps its original body
+        while the remaining entries are still compressed, preserving already
+        retrieved search results.
+        """
+        assert self.context is not None
+        # Guard: when the compressor job is missing from the active config
+        # (e.g. a benchmark config that forgot to define it), degrade
+        # gracefully to the no-compression behavior instead of raising
+        # "Job compressor not found" from run_job below.
+        # Skipped when there is no app_context (e.g. unit tests that mock
+        # run_job directly), so the mock can still drive compression.
+        if self.app_context is not None and self.get_job("compressor") is None:
+            self.logger.warning(
+                f"[{self.name}] compressor job not found in config; "
+                "skipping session chunk compression (degrading to uncompressed behavior)",
+            )
+            return
+        search_cfg: dict = self.context.get("_search") or {}
+        query_type = str(search_cfg.get("type") or "query-aware").strip().lower()
+        if query_type == "query-independent":
+            queries: list[str] = []
+        else:
+            queries = [str(q).strip() for q in (search_cfg.get("queries") or []) if str(q).strip()]
+            if query and query not in queries:
+                queries.append(query)
+
+        async def compress(entry: dict[str, str]) -> None:
+            path = entry.get("path", "")
+            body = (entry.get("body", "") or "").strip()
+            if not body:
+                return
+            try:
+                response = await self.run_job("compressor", text=body, queries=queries)
+            except Exception as exc:  # pylint: disable=broad-except
+                self.logger.warning(
+                    f"[{self.name}] session body compression raised path={path!r} " f"error={exc!r}; keeping original",
+                )
+                return
+            compressed = str(response.answer or "").strip()
+            if not response.success or not compressed:
+                self.logger.warning(
+                    f"[{self.name}] session body compression failed path={path!r} "
+                    f"success={response.success} answer={compressed[:100]!r}; keeping original",
+                )
+                return
+            if len(compressed) > len(body):
+                self.logger.info(
+                    f"[{self.name}] compressed body longer than original "
+                    f"({len(compressed)} > {len(body)}) path={path!r}; keeping original",
+                )
+                return
+            entry["body"] = f"compressed session chunk:\n{compressed}"
+
+        targets = [e for e in entries if is_session_path(e.get("path", ""), session_dir)]
+        if not targets:
+            return
+        self.logger.info(f"[{self.name}] compressing {len(targets)} session entries with {len(queries)} queries")
+        await asyncio.gather(*(compress(entry) for entry in targets))
--- a/reme/config/beam.yaml
+++ b/reme/config/beam.yaml
@ -1,635 +0,0 @@
-# BEAM benchmark config — based on longmemeval.yaml (split)
-# All background/cron jobs are converted to base (manually callable).
-# Use with: resolve_app_config(config="beam.yaml", ...)
-
-service:
-  backend: http
-
-jobs:
-  # ── BEAM agentic answer (ReAct agent + search tool) ──
-  agentic_answer:
-    backend: base
-    description: "BEAM agentic answer job (ReAct agent with search tool)"
-    watch_dirs: []
-    watch_suffixes: []
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-          description: "The query to ask"
-        query_time:
-          type: string
-          description: "ISO timestamp representing the query time"
-          default: ""
-      required:
-        - query
-    steps:
-      - backend: beam_agentic_answer_step
-        agent_wrapper: bench
-
-  # ── BEAM rubric-based LLM-as-Judge ──
-  answer_judge:
-    backend: base
-    description: "BEAM rubric-based LLM-as-Judge: evaluate response against rubric criteria"
-    watch_dirs: []
-    watch_suffixes: []
-    parameters:
-      type: object
-      properties:
-        llm_response:
-          type: string
-          description: "The model's response to evaluate"
-        rubric:
-          type: array
-          description: "List of rubric criteria to check"
-          items:
-            type: string
-        probing_question:
-          type: string
-          description: "The original probing question"
-          default: ""
-        question_type:
-          type: string
-          description: "BEAM question type (e.g. event_ordering)"
-          default: ""
-      required:
-        - llm_response
-        - rubric
-    steps:
-      - backend: beam_rubric_judge_step
-        agent_wrapper: judge
-
-  # ── Manual index update (replaces index_update_loop background) ──
-  index_update:
-    backend: base
-    description: "Manually trigger incremental index update for watched dirs."
-    watch_dirs: [daily_dir, digest_dir, session_dir/dialog]
-    watch_suffixes: [md, jsonl]
-    parameters:
-      type: object
-      properties: {}
-    steps:
-      - backend: init_changes_step
-        monitor_type: file_store
-        monitor_name: default
-        dispatch_steps: [update_index_step]
-
-  # ── Manual digest catalog update (replaces digest_watch_loop background) ──
-  digest_update:
-    backend: base
-    description: "Manually trigger digest catalog update."
-    watch_dirs: [daily_dir, digest_dir]
-    watch_suffixes: [md]
-    parameters:
-      type: object
-      properties: {}
-    steps:
-      - backend: init_changes_step
-        monitor_type: file_catalog
-        monitor_name: digest
-        dispatch_steps:
-          - backend: update_catalog_step
-            file_catalog: digest
-          - backend: log_changes_step
-
-  # ── Auto dream (same as default.yaml auto_dream, base mode) ──
-  # auto_dream:
-  #   backend: base
-  #   description: "Auto-dream: scan today's day-index and daily notes, globally extract merged units/topics, integrate digest units, write interests.yaml, and persist the dream catalog."
-  #   parameters:
-  #     type: object
-  #     properties:
-  #       date:
-  #         type: string
-  #         description: "YYYY-MM-DD to scan; defaults to today in the dreamer's timezone"
-  #         default: ""
-  #       hint:
-  #         type: string
-  #         description: "caller guidance passed through to dream extract/integrate"
-  #         default: ""
-  #       scan_days:
-  #         type: integer
-  #         description: "number of recent daily directories to scan, ending at date"
-  #         default: 2
-  #       max_units:
-  #         type: integer
-  #         description: "maximum number of extracted memory units"
-  #         default: 5
-  #       topic_count:
-  #         type: integer
-  #         description: "maximum number of final daily interest topics"
-  #         default: 3
-  #       topic_diversity_days:
-  #         type: integer
-  #         description: "number of previous interests.yaml days to avoid repeating"
-  #         default: 7
-  #   steps:
-  #     - backend: dream_extract_step
-  #       file_catalog: dream
-  #       topic_session_id: interests
-  #       scan_days: 2
-  #       max_units: 5
-  #     - backend: dream_integrate_step
-  #     - backend: dream_topics_step
-  #       topic_count: 3
-  #       topic_diversity_days: 7
-  #     - backend: dream_finish_step
-  #       file_catalog: dream
-
-  # ── Auto memory (same as default.yaml) ──
-  auto_memory:
-    backend: base
-    description: "Auto-memory: record conversation facts into a daily note"
-    parameters:
-      type: object
-      properties:
-        messages:
-          type: array
-          description: "messages"
-          items:
-            type: object
-        session_id:
-          type: string
-          description: "source conversation session identifier"
-          default: ""
-        memory_hint:
-          type: string
-          description: "optional hint"
-        date:
-          type: string
-          description: "YYYY-MM-DD daily note date; empty = infer from message timestamps or today"
-          default: ""
-      required:
-        - messages
-    steps:
-      - backend: beam_auto_memory_step
-        # Long sessions are split into turn-aligned segments and fed to the
-        # agent incrementally; each segment holds at most this many words.
-        # ("segment" here is a slice of the message list, unrelated to file
-        # chunking in the index.) <= 0 disables splitting.
-        max_segment_words: 10000
-
-  # ── Text compression (direct LLM call, no agent) ──
-  compressor:
-    backend: base
-    description: "Compress text via a direct LLM call, optionally guided by queries as relevance filter"
-    parameters:
-      type: object
-      properties:
-        text:
-          type: string
-          description: "the text to compress"
-        queries:
-          type: array
-          description: "optional list of queries; content potentially relevant to any query is kept, content certainly irrelevant to all queries may be dropped"
-          items:
-            type: string
-          default: []
-      required:
-        - text
-    steps:
-      - backend: compressor_step
-        as_llm: compressor
-
-  # ── Reindex (derived search indexes only) ──
-  reindex:
-    backend: base
-    description: "rebuild BM25 and/or embedding indexes from current file_chunks"
-    parameters:
-      type: object
-      properties:
-        scope:
-          type: string
-          enum: [all, bm25, embedding]
-          default: all
-    steps:
-      - backend: reindex_step
-
-  # ── Search ──
-  #   start_date:
-        #   type: string
-        #   description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
-        # end_date:
-        #   type: string
-        #   description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
-
-  search:
-    backend: base
-    description: "Hybrid workspace search (vector + BM25, RRF-fused) with deduplication."
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-          description: "search query"
-        start_date:
-          type: string
-          description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
-        end_date:
-          type: string
-          description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
-        # vector_weight:
-        #   type: number
-        #   description: >-
-        #     Optional weight balancing vector similarity vs BM25 keyword matching in the
-        #     RRF fusion. Recommended value is 0.7, which provides a good balance between
-        #     semantic (vector) similarity and lexical (BM25) matching. Values close to 0
-        #     emphasize BM25 keyword matching, values close to 1 emphasize vector semantic
-        #     similarity.
-      required:
-        - query
-    steps:
-      - backend: search_v2_step
-        vector_weight: 0.7
-        candidate_multiplier: 5.0
-        expand_links: false
-        max_links_per_direction: 10
-
-  add_draft:
-    backend: base
-    description: "Append text to the current draft list."
-    parameters:
-      type: object
-      properties:
-        text:
-          type: string
-          description: "draft text to append"
-      required:
-        - text
-    steps:
-      - backend: add_draft_step
-
-  read_all_draft:
-    backend: base
-    description: "Read all draft text previously appended in the current tool context."
-    parameters:
-      type: object
-      properties: { }
-    steps:
-      - backend: read_all_draft_step
-
-  python_execute:
-    backend: base
-    description: "Execute Python code and return printed stdout."
-    parameters:
-      type: object
-      properties:
-        code:
-          type: string
-          description: "Python code to execute. Print the final result to stdout."
-        timeout:
-          type: number
-          description: "Execution timeout in seconds; defaults to 60."
-      required:
-        - code
-    steps:
-      - backend: python_execute_step
-
-  # ── File I/O jobs (needed by auto_memory agent tools) ──
-  daily_list:
-    backend: base
-    description: "List notes under a single day."
-    parameters:
-      type: object
-      properties:
-        date:
-          type: string
-          description: "YYYY-MM-DD; empty = today"
-          default: ""
-    steps:
-      - backend: daily_list_step
-
-  daily_reindex:
-    backend: base
-    description: "Rebuild the day-index page daily/<date>.md."
-    parameters:
-      type: object
-      properties:
-        date:
-          type: string
-          description: "YYYY-MM-DD; empty = today"
-          default: ""
-    steps:
-      - backend: daily_reindex_step
-
-  frontmatter_update:
-    backend: base
-    description: "Merge key-values into a file's frontmatter."
-    parameters:
-      type: object
-      properties:
-        path:
-          type: string
-          description: "workspace-relative path"
-        metadata:
-          type: object
-          description: "key-values to merge"
-      required:
-        - path
-        - metadata
-    steps:
-      - backend: frontmatter_update_step
-
-  move:
-    backend: base
-    description: "Move / rename a workspace file."
-    parameters:
-      type: object
-      properties:
-        src_path:
-          type: string
-          description: "workspace-relative source"
-        dst_path:
-          type: string
-          description: "workspace-relative destination"
-        overwrite:
-          type: boolean
-          default: false
-        retarget:
-          type: boolean
-          default: true
-      required:
-        - src_path
-        - dst_path
-    steps:
-      - backend: move_step
-
-  read:
-    backend: base
-    description: "Read a markdown file under the workspace."
-    parameters:
-      type: object
-      properties:
-        path:
-          type: string
-          description: "workspace-relative path; markdown only"
-        start_line:
-          type: integer
-        end_line:
-          type: integer
-      required:
-        - path
-    steps:
-      - backend: read_step
-        with_neighbors: false
-        max_neighbors_per_direction: 10
-
-  write:
-    backend: base
-    description: "Write a markdown file."
-    parameters:
-      type: object
-      properties:
-        path:
-          type: string
-        name:
-          type: string
-        description:
-          type: string
-        content:
-          type: string
-        metadata:
-          type: object
-      required:
-        - path
-        - name
-        - description
-        - content
-    steps:
-      - backend: write_step
-
-  daily_write:
-    backend: base
-    description: "Write a daily markdown note."
-    parameters:
-      type: object
-      properties:
-        name:
-          type: string
-        description:
-          type: string
-        session_id:
-          type: string
-        content:
-          type: string
-        date:
-          type: string
-          default: ""
-        metadata:
-          type: object
-      required:
-        - name
-        - description
-        - session_id
-        - content
-    steps:
-      - backend: daily_write_step
-
-  edit:
-    backend: base
-    description: "Find-and-replace in a markdown file."
-    parameters:
-      type: object
-      properties:
-        path:
-          type: string
-        old:
-          type: string
-        new:
-          type: string
-          default: ""
-      required:
-        - path
-        - old
-        - new
-    steps:
-      - backend: edit_step
-
-  frontmatter_read:
-    backend: base
-    description: "Read a file's frontmatter as a dict."
-    parameters:
-      type: object
-      properties:
-        path:
-          type: string
-      required:
-        - path
-    steps:
-      - backend: frontmatter_read_step
-
-  node_search:
-    backend: base
-    description: "Digest node recall."
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-        limit:
-          type: integer
-          default: 20
-      required:
-        - query
-    steps:
-      - backend: node_search_step
-        vector_weight: 0.7
-        candidate_multiplier: 5.0
-
-components:
-  tokenizer:
-    default:
-      backend: regex
-
-  as_embedding:
-    default:
-      backend: ${EMBEDDING_BACKEND:-openai}
-      model: ${EMBEDDING_MODEL_NAME:-text-embedding-v4}
-      credential:
-        api_key: ${EMBEDDING_API_KEY:-}
-        base_url: ${EMBEDDING_BASE_URL:-https://dashscope.aliyuncs.com/compatible-mode/v1}
-      dimensions: 1024
-
-  embedding_store:
-    default:
-      backend: local
-      as_embedding: default
-
-  as_llm:
-    default:
-      backend: ${LLM_BACKEND:-openai}
-      model: ${LLM_MODEL_NAME:-qwen3.6-flash}
-      stream: true
-      context_size: 200000
-      max_retries: 5
-      retry_delay: 5.0
-      credential:
-        api_key: ${LLM_API_KEY:-}
-        base_url: ${LLM_BASE_URL:-}
-      parameters:
-        max_tokens: 65536
-        thinking_enable: false
-    judge:
-      backend: ${LLM_BACKEND:-openai}
-      model: ${JUDGE_MODEL_NAME:-qwen3.7-max}
-      stream: false
-      context_size: 200000
-      max_retries: 5
-      retry_delay: 5.0
-      credential:
-        api_key: ${LLM_API_KEY:-}
-        base_url: ${LLM_BASE_URL:-}
-      parameters:
-        max_tokens: 65536
-        thinking_enable: false
-    bench:
-      backend: ${LLM_BACKEND:-openai}
-      model: ${BENCH_MODEL_NAME:-qwen3.7-max}
-      stream: true
-      context_size: 400000
-      max_retries: 5
-      retry_delay: 5.0
-      credential:
-        api_key: ${LLM_API_KEY:-}
-        base_url: ${LLM_BASE_URL:-}
-      parameters:
-        max_tokens: 65536
-        thinking_enable: true
-    compressor:
-      backend: ${LLM_BACKEND:-openai}
-      model: ${LLM_MODEL_NAME:-qwen3.6-flash}
-      stream: false
-      context_size: 200000
-      max_retries: 5
-      retry_delay: 5.0
-      credential:
-        api_key: ${LLM_API_KEY:-}
-        base_url: ${LLM_BASE_URL:-}
-      parameters:
-        max_tokens: 65536
-        thinking_enable: false
-
-  agent_wrapper:
-    default:
-      backend: agentscope
-      as_llm: default
-      permission_mode: bypass
-      react_config:
-        max_iters: 30
-      context_config:
-        trigger_ratio: 0.8
-        reserve_ratio: 0.1
-        tool_result_limit: 50000
-      model_config:
-        max_retries: 1
-    judge:
-      backend: agentscope
-      as_llm: judge
-      permission_mode: bypass
-      react_config:
-        max_iters: 1
-      context_config:
-        trigger_ratio: 0.8
-        reserve_ratio: 0.1
-        tool_result_limit: 50000
-      model_config:
-        max_retries: 1
-    bench:
-      backend: agentscope
-      as_llm: bench
-      permission_mode: bypass
-      react_config:
-        max_iters: 30
-      context_config:
-        trigger_ratio: 0.8
-        reserve_ratio: 0.1
-        tool_result_limit: 50000
-      model_config:
-        max_retries: 1
-
-  file_graph:
-    default:
-      backend: local
-
-  file_catalog:
-    default:
-      backend: local
-    resource:
-      backend: local
-    digest:
-      backend: local
-    dream:
-      backend: local
-
-  file_chunker:
-    markdown:
-      backend: markdown
-      supported_extensions: [ "md" ]
-      embed_toc: true
-      max_ast_sections: 100
-      include_frontmatter_in_metadata: false
-      include_frontmatter_keys_in_metadata: []  # empty = all non-empty frontmatter keys
-    json:
-      backend: json
-      supported_extensions: [ "json" ]
-    jsonl:
-      backend: jsonl
-      supported_extensions: [ "jsonl" ]  # noqa: keep #314 chunker scope intact after #325
-      max_chars: 4000
-    default:
-      backend: default
-      supported_extensions: ["txt","log"]
-
-  keyword_index:
-    default:
-      backend: bm25
-      tokenizer: default
-
-  file_store:
-    default:
-      backend: local
-      store_name: local
-      embedding_store: default
-      keyword_index: default
-      file_graph: default
--- a/reme/config/benchmark.yaml
+++ b/reme/config/benchmark.yaml
@ -1,63 +1,11 @@
-# LongMemEval benchmark config — based on longmemeval.yaml (split)
-# All background/cron jobs are converted to base (manually callable).
-# Use with: resolve_app_config(config="lme.yaml", ...)
+# Shared benchmark application preset; independent of the default service config.
+# Background and cron jobs are omitted in favor of manually callable base jobs.
+# Benchmark plugins contribute their own jobs through plugin.yaml.

 service:
  backend: http

 jobs:
-  # ── LongMemEval agentic answer (ReAct agent + search tool) ──
-  agentic_answer:
-    backend: base
-    description: "LongMemEval agentic answer job (ReAct agent with search tool)"
-    watch_dirs: []
-    watch_suffixes: []
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-          description: "The query to ask"
-        query_time:
-          type: string
-          description: "ISO timestamp representing the query time"
-          default: ""
-      required:
-        - query
-    steps:
-      - backend: lme_agentic_answer_step
-        agent_wrapper: bench
-
-  # ── LLM-as-Judge for evaluating answers ──
-  answer_judge:
-    backend: base
-    description: "LLM-as-Judge: evaluate agent answer against golden answer"
-    watch_dirs: []
-    watch_suffixes: []
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-          description: "The question being asked"
-        agent_answer:
-          type: string
-          description: "The model's answer to evaluate"
-        golden_answer:
-          type: string
-          description: "The correct/golden answer"
-        question_type:
-          type: string
-          description: "The question type for prompt selection"
-          default: ""
-      required:
-        - query
-        - agent_answer
-        - golden_answer
-    steps:
-      - backend: lme_answer_judge_step
-        agent_wrapper: judge
-
  # ── Manual index update (replaces index_update_loop background) ──
  index_update:
    backend: base
@ -135,34 +83,6 @@ jobs:
  #     - backend: dream_finish_step
  #       file_catalog: dream

-  # ── Auto memory (same as default.yaml) ──
-  auto_memory:
-    backend: base
-    description: "Auto-memory: record conversation facts into a daily note"
-    parameters:
-      type: object
-      properties:
-        messages:
-          type: array
-          description: "messages"
-          items:
-            type: object
-        session_id:
-          type: string
-          description: "source conversation session identifier"
-          default: ""
-        memory_hint:
-          type: string
-          description: "optional hint"
-        date:
-          type: string
-          description: "YYYY-MM-DD daily note date; empty = infer from message timestamps or today"
-          default: ""
-      required:
-        - messages
-    steps:
-      - backend: lme_auto_memory_step
-
  # ── Text compression (direct LLM call, no agent) ──
  compressor:
    backend: base
@ -199,46 +119,6 @@ jobs:
    steps:
      - backend: reindex_step

-  # ── Search ──
-  #   start_date:
-        #   type: string
-        #   description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
-        # end_date:
-        #   type: string
-        #   description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
-
-  search:
-    backend: base
-    description: "Hybrid workspace search (vector + BM25, RRF-fused) with deduplication."
-    parameters:
-      type: object
-      properties:
-        query:
-          type: string
-          description: "search query"
-        start_date:
-          type: string
-          description: "optional inclusive start date filter (YYYY-MM-DD); results earlier than this date are excluded"
-        end_date:
-          type: string
-          description: "optional inclusive end date filter (YYYY-MM-DD); results later than this date are excluded"
-        # vector_weight:
-        #   type: number
-        #   description: >-
-        #     Optional weight balancing vector similarity vs BM25 keyword matching in the
-        #     RRF fusion. Recommended value is 0.7, which provides a good balance between
-        #     semantic (vector) similarity and lexical (BM25) matching. Values close to 0
-        #     emphasize BM25 keyword matching, values close to 1 emphasize vector semantic
-        #     similarity.
-      required:
-        - query
-    steps:
-      - backend: search_v2_step
-        vector_weight: 0.7
-        candidate_multiplier: 5.0
-        expand_links: false
-        max_links_per_direction: 10
-
  add_draft:
    backend: base
    description: "Append text to the current draft list."
@ -326,7 +206,7 @@ jobs:

  move:
    backend: base
-    description: "Move / rename a workspace file."
+    description: "Move / rename a workspace file; rewrites inbound wikilinks by default."
    parameters:
      type: object
      properties:
@ -338,9 +218,11 @@ jobs:
          description: "workspace-relative destination"
        overwrite:
          type: boolean
+          description: "overwrite if dst exists"
          default: false
        retarget:
          type: boolean
+          description: "rewrite [[src]] → [[dst]] across the workspace"
          default: true
      required:
        - src_path
@ -359,8 +241,10 @@ jobs:
          description: "workspace-relative path; markdown only"
        start_line:
          type: integer
+          description: "first line (1-based, inclusive)"
        end_line:
          type: integer
+          description: "last line (1-based, inclusive)"
      required:
        - path
    steps:
@ -370,20 +254,25 @@ jobs:

  write:
    backend: base
-    description: "Write a markdown file."
+    description: "Write a markdown file (create or overwrite) with name/description frontmatter."
    parameters:
      type: object
      properties:
        path:
          type: string
+          description: "workspace-relative path; markdown only"
        name:
          type: string
+          description: "frontmatter name"
        description:
          type: string
+          description: "frontmatter description"
        content:
          type: string
+          description: "body"
        metadata:
          type: object
+          description: "Optional extra frontmatter fields (md only)."
      required:
        - path
        - name
@ -394,23 +283,29 @@ jobs:

  daily_write:
    backend: base
-    description: "Write a daily markdown note."
+    description: "Write a daily markdown note with conversation source frontmatter."
    parameters:
      type: object
      properties:
        name:
          type: string
+          description: "daily note filename stem and frontmatter name"
        description:
          type: string
+          description: "frontmatter description"
        session_id:
          type: string
+          description: "source conversation session identifier"
        content:
          type: string
+          description: "body"
        date:
          type: string
+          description: "YYYY-MM-DD daily note date; empty = today"
          default: ""
        metadata:
          type: object
+          description: "Optional extra frontmatter fields."
      required:
        - name
        - description
@ -421,16 +316,19 @@ jobs:

  edit:
    backend: base
-    description: "Find-and-replace in a markdown file."
+    description: "Find-and-replace in a markdown file (all occurrences)."
    parameters:
      type: object
      properties:
        path:
          type: string
+          description: "workspace-relative path"
        old:
          type: string
+          description: "text to find"
        new:
          type: string
+          description: "replacement"
          default: ""
      required:
        - path
@ -447,6 +345,7 @@ jobs:
      properties:
        path:
          type: string
+          description: "workspace-relative path"
      required:
        - path
    steps:
@ -454,14 +353,16 @@ jobs:

  node_search:
    backend: base
-    description: "Digest node recall."
+    description: "Digest node recall — given a candidate abstraction's name+description, surface existing digest nodes similar enough to either dedup against or link to as related."
    parameters:
      type: object
      properties:
        query:
          type: string
+          description: "search query"
        limit:
          type: integer
+          description: "max digest nodes to return"
          default: 20
      required:
        - query
@ -495,7 +396,6 @@ components:
      model: ${LLM_MODEL_NAME:-qwen3.6-flash}
      stream: true
      context_size: 200000
-      max_retries: 3
      retry_delay: 5.0
      credential:
        api_key: ${LLM_API_KEY:-}
--- a/reme/steps/benchmark/init.py
+++ b/reme/steps/benchmark/init.py
@ -1,17 +1,5 @@
-"""Benchmark steps."""
+"""Shared benchmark steps; concrete implementations live in plugins."""

-from . import base, lme, beam
-from .base import BaseAgenticAnswerStep
-from .lme import LmeAgenticAnswerStep, LmeAnswerJudgeStep
-from .beam import BeamAgenticAnswerStep, BeamRubricJudgeStep
+from .base_agentic_answer import BaseAgenticAnswerStep

-__all__ = [
-    "BaseAgenticAnswerStep",
-    "LmeAgenticAnswerStep",
-    "LmeAnswerJudgeStep",
-    "BeamAgenticAnswerStep",
-    "BeamRubricJudgeStep",
-    "base",
-    "lme",
-    "beam",
-]
+__all__ = ["BaseAgenticAnswerStep"]
--- a/reme/steps/benchmark/base/init.py
+++ b/reme/steps/benchmark/base/init.py
@ -1,7 +0,0 @@
-"""Shared base classes for benchmark steps."""
-
-from .agentic_answer import BaseAgenticAnswerStep
-
-__all__ = [
-    "BaseAgenticAnswerStep",
-]
--- a/reme/steps/benchmark/base_agentic_answer.py
+++ b/reme/steps/benchmark/base_agentic_answer.py
@ -2,14 +2,14 @@

 import os

-from ...base_step import BaseStep
-from ...index._dedup import _ToolContextDedupMixin
-from ....enumeration import ChunkEnum
-from ....utils.counter import global_counter_inc
+from ..base_step import BaseStep
+from ..index._dedup import _ToolContextDedupMixin
+from ...enumeration import ChunkEnum
+from ...utils.counter import global_counter_inc


 class BaseAgenticAnswerStep(BaseStep):
-    """Base ReAct-agent answer step shared by BEAM and LongMemEval benchmarks.
+    """ReAct-agent answer implementation shared by benchmark plugins.

    Subclasses only need to set:
        TOOL_CONTEXT_PREFIX (str): prefix used to build the unique tool_context_id.
@ -18,7 +18,7 @@ class BaseAgenticAnswerStep(BaseStep):
            tool call via ``injected_job_kwargs``; override the attribute or the
            ``_injected_job_kwargs`` hook to customize.

-    And apply their own ``@R.register(...)`` decorator and docstring.
+    Concrete subclasses are registered by the plugin manifest.

    Inputs (from RuntimeContext):
        query       (str, required): The question to answer.
@ -43,7 +43,7 @@ class BaseAgenticAnswerStep(BaseStep):
        ``INJECTED_JOB_KWARGS`` with per-request values derived from ``query``.

        When the runtime context carries a truthy ``compress_session`` flag,
-        session-transcript compression is enabled in ``search_v2_step`` by
+        session-transcript compression is enabled in the benchmark plugin's search Step by
        injecting a ``_search._compress.session`` marker plus the current
        ``query`` as the query-aware relevance filter. Default (falsy) leaves
        session chunks uncompressed.
--- a/reme/steps/index/init.py
+++ b/reme/steps/index/init.py
@ -12,7 +12,6 @@ from .init_changes import InitChangesStep
 from .optimize_index import OptimizeIndexStep
 from .reindex import ReindexStep
 from .search import SearchStep
-from .search_v2 import SearchV2Step
 from .traverse import TraverseStep
 from .update_changes import ChangeApplyStep, UpdateCatalogStep, UpdateIndexStep
 from .vector_search import VectorSearchStep
@ -42,7 +41,6 @@ __all__ = [
    "ReindexStep",
    "OptimizeIndexStep",
    "SearchStep",
-    "SearchV2Step",
    "TraverseStep",
    "UpdateCatalogStep",
    "UpdateIndexStep",
--- a/tests/unit/test_beam_auto_memory.py
+++ b/tests/unit/test_beam_auto_memory.py
@ -1,128 +0,0 @@
-"""Unit tests for BEAM auto-memory segmented ingestion and line numbering."""
-
-# pylint: disable=missing-class-docstring,missing-function-docstring,protected-access
-
-from agentscope.message import Msg
-
-from reme.components.runtime_context import RuntimeContext
-from reme.steps.benchmark.beam.auto_memory import (
-    BeamAutoMemoryStep,
-    split_turn_segments,
-)
-
-
-def _msg(role: str, words: int, created_at: str = "2024-03-01T09:00:00") -> Msg:
-    return Msg(
-        name=role,
-        role=role,
-        content=[{"type": "text", "text": " ".join(["w"] * words)}],
-        created_at=created_at,
-    )
-
-
-def _dialog(n_turns: int, words_per_msg: int) -> list[Msg]:
-    messages: list[Msg] = []
-    for _ in range(n_turns):
-        messages.append(_msg("user", words_per_msg))
-        messages.append(_msg("assistant", words_per_msg))
-    return messages
-
-
-class TestSplitTurnSegments:
-    def test_empty(self):
-        assert not split_turn_segments([], 100)
-
-    def test_disabled_returns_single_segment(self):
-        messages = _dialog(3, 10)
-        segments = split_turn_segments(messages, 0)
-        assert len(segments) == 1
-        assert segments[0] == (0, messages)
-
-    def test_under_limit_single_segment(self):
-        messages = _dialog(3, 10)  # 60 words total
-        segments = split_turn_segments(messages, 100)
-        assert len(segments) == 1
-        assert segments[0][0] == 0
-        assert segments[0][1] == messages
-
-    def test_splits_at_turn_boundaries(self):
-        # 4 turns x 20 words each; limit 40 -> 2 turns per segment
-        messages = _dialog(4, 10)
-        segments = split_turn_segments(messages, 40)
-        assert len(segments) == 2
-        offsets = [offset for offset, _ in segments]
-        assert offsets == [0, 4]
-        # Every segment starts with a user message and ends with an assistant
-        for _, segment in segments:
-            assert segment[0].role == "user"
-            assert segment[-1].role == "assistant"
-        # No message lost or duplicated, order preserved
-        flattened = [m for _, segment in segments for m in segment]
-        assert flattened == messages
-
-    def test_never_splits_inside_a_turn(self):
-        # One turn alone exceeds the limit -> becomes its own oversized segment
-        messages = [
-            _msg("user", 5),
-            _msg("assistant", 5),
-            _msg("user", 50),
-            _msg("assistant", 50),  # 100-word turn > limit 60
-            _msg("user", 5),
-            _msg("assistant", 5),
-        ]
-        segments = split_turn_segments(messages, 60)
-        assert [offset for offset, _ in segments] == [0, 2, 4]
-        assert [len(segment) for _, segment in segments] == [2, 2, 2]
-
-    def test_offsets_are_original_indices(self):
-        messages = _dialog(5, 30)  # 60 words per turn
-        segments = split_turn_segments(messages, 120)
-        # 2 turns per segment -> offsets 0, 4, 8
-        assert [offset for offset, _ in segments] == [0, 4, 8]
-        for offset, segment in segments:
-            for i, msg in enumerate(segment):
-                assert msg is messages[offset + i]
-
-    def test_multi_assistant_turn_stays_together(self):
-        messages = [
-            _msg("user", 10),
-            _msg("assistant", 10),
-            _msg("assistant", 10),
-            _msg("user", 10),
-            _msg("assistant", 10),
-        ]
-        segments = split_turn_segments(messages, 30)
-        assert [offset for offset, _ in segments] == [0, 3]
-        assert len(segments[0][1]) == 3
-
-
-class TestFormatHistoryLineNumbers:
-    def _step(self, offset: int) -> BeamAutoMemoryStep:
-        step = BeamAutoMemoryStep(name="beam_auto_memory_step", backend="beam_auto_memory_step")
-        step.context = RuntimeContext(
-            session_id="beam_1M_1_batch1",
-            beam_line_offset=offset,
-        )
-        return step
-
-    def test_numbers_start_at_one_without_offset(self):
-        step = self._step(0)
-        history = step._format_history(_dialog(2, 3))
-        assert "[L1 | user @" in history
-        assert "[L4 | assistant @" in history
-        assert "lines 1-4" in history
-        assert "session/dialog/beam_1M_1_batch1.jsonl" in history
-
-    def test_numbers_use_original_file_offset(self):
-        step = self._step(40)
-        history = step._format_history(_dialog(2, 3))
-        assert "[L41 | user @" in history
-        assert "[L44 | assistant @" in history
-        assert "lines 41-44" in history
-        assert "[L1 " not in history
-
-    def test_build_messages_passes_msg_objects_through(self):
-        step = self._step(0)
-        messages = _dialog(2, 3)
-        rebuilt = step._build_messages(messages)
-        assert [m.id for m in rebuilt] == [m.id for m in messages]
--- a/tests/unit/test_injected_job_kwargs.py
+++ b/tests/unit/test_injected_job_kwargs.py
@ -230,7 +230,7 @@ async def test_auto_memory_update_scopes_tools_to_exact_note_path(tmp_path, monk


 def test_auto_memory_keeps_original_tool_names():
-    """BEAM/LME configs define only the original jobs; no *_daily variants exist."""
+    """Core auto-memory uses the original file tool names."""
    step = AutoMemoryStep(name="auto_memory")
    assert step.create_tools == ["daily_write"]
    assert step.update_tools == ["read", "edit", "frontmatter_update", "write"]
@ -240,11 +240,7 @@ def test_auto_memory_create_prompts_match_upstream_date_arguments():
    """Auto-memory prompts keep the upstream model-supplied date argument."""
    from pathlib import Path

-    prompt_files = (
-        Path("reme/steps/evolve/auto_memory.yaml"),
-        Path("reme/steps/benchmark/beam/auto_memory.yaml"),
-        Path("reme/steps/benchmark/lme/auto_memory.yaml"),
-    )
+    prompt_files = (Path("reme/steps/evolve/auto_memory.yaml"),)
    for prompt_file in prompt_files:
        content = prompt_file.read_text(encoding="utf-8")
        assert "date={today}" in content or "`date`: {today}" in content or "`date`：{today}" in content
@ -253,7 +249,7 @@ def test_auto_memory_create_prompts_match_upstream_date_arguments():
 def test_configs_define_original_jobs_without_daily_variants():
    from reme.config import resolve_app_config

-    for config_name in ("default", "lme", "beam"):
+    for config_name in ("default",):
        config = resolve_app_config(config=config_name, log_config=False)
        jobs = config["jobs"]
        for name in ("read", "edit", "write", "frontmatter_update", "daily_write"):
--- a/tests/unit/test_plugin.py
+++ b/tests/unit/test_plugin.py
@ -43,6 +43,22 @@ def _set_entry_points(monkeypatch, *entries):
    monkeypatch.setattr("reme.entry_point.metadata.entry_points", lambda: _FakeEntryPoints(entries))


+@pytest.mark.parametrize("name", ["lme", "beam"])
+def test_shared_benchmark_preset_is_builtin_but_plugin_aliases_and_backends_are_not(monkeypatch, name):
+    _set_entry_points(monkeypatch)
+    benchmark = _load_config("benchmark")
+    assert {"index_update", "digest_update", "read", "write"} <= benchmark["jobs"].keys()
+    assert {"search", "auto_memory", "agentic_answer", "answer_judge"}.isdisjoint(benchmark["jobs"])
+    for alias in (name, f"{name}.yaml"):
+        with pytest.raises(FileNotFoundError, match="Config file not found"):
+            _load_config(alias)
+    assert R.get(ComponentEnum.STEP, f"{name}_auto_memory_step") is None
+    assert R.get(ComponentEnum.STEP, f"{name}_agentic_answer_step") is None
+    assert R.get(ComponentEnum.STEP, f"{name}_search_v2_step") is None
+    judge = "lme_answer_judge_step" if name == "lme" else "beam_rubric_judge_step"
+    assert R.get(ComponentEnum.STEP, judge) is None
+
+
 def test_plugin_application_defaults_are_below_application_config():
    manager = PluginManager(
        [
--- a/tests/unit/test_search_step.py
+++ b/tests/unit/test_search_step.py
@ -1,6 +1,8 @@
-"""Unit tests for SearchV2Step without embedding or LLM dependencies."""
+"""Unit tests for workspace search Steps without embedding or LLM dependencies."""

 import asyncio
+import importlib.util
+from pathlib import Path

 from agentscope.message import Msg

@ -14,11 +16,17 @@ from reme.steps.index import (
    Bm25SearchStep,
    ReadAllDraftStep,
    SearchStep,
-    SearchV2Step,
    VectorSearchStep,
 )
 from reme.steps.index._source_format import ALL_RETURNED_MESSAGE, NO_RESULTS_MESSAGE

+_SEARCH_V2_PATH = Path(__file__).parents[2] / "plugins" / "beam" / "src" / "reme_beam" / "search_v2.py"
+_SEARCH_V2_SPEC = importlib.util.spec_from_file_location("reme_beam.search_v2", _SEARCH_V2_PATH)
+assert _SEARCH_V2_SPEC is not None and _SEARCH_V2_SPEC.loader is not None
+_SEARCH_V2_MODULE = importlib.util.module_from_spec(_SEARCH_V2_SPEC)
+_SEARCH_V2_SPEC.loader.exec_module(_SEARCH_V2_MODULE)
+SearchV2Step = _SEARCH_V2_MODULE.SearchV2Step
+

 class FakeSearchStore(BaseFileStore):
    """Minimal file_store for SearchV2Step: static search results and empty graph links."""