AMAP-ML · lerogo · Aug 14, 2026 · Aug 14, 2026
diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 
 ### Loop Engineering for Computer-Use Agents
 
-**Give Claude Code or Codex a goal once. Keep it working across desktop apps and the terminal for dozens of hours.**
+**Give Claude Code, Codex, or DeepSeek Harness a goal once. Keep it working across desktop apps and the terminal for dozens of hours.**
 
 **Plan → act → verify → checkpoint or recover → repeat — until the work is actually done.**
 
@@ -18,7 +18,7 @@
 </p>
 
 [![Python](https://img.shields.io/badge/python-≥3.10-blue?logo=python&logoColor=white)](https://www.python.org/)
-[![Agents](https://img.shields.io/badge/backends-Claude%20Code%20|%20Codex-8A2BE2)](#any-model-any-agent-backend)
+[![Agents](https://img.shields.io/badge/backends-Claude%20Code%20|%20Codex%20|%20DeepSeek-8A2BE2)](#any-model-any-agent-backend)
 [![Benchmarks](https://img.shields.io/badge/benchmarks-WeaveBench%20|%20OSWorld%202.0%20|%20Terminal--Bench%202.1-orange)](#hundreds-of-real-tasks-measured-gains)
 
 [Usage](#one-command-full-visibility) · [The Loop](#loop-engineering-for-real-computer-environments) · [Computer Use](#desktop-apps-and-cli-one-continuous-task) · [Results](#hundreds-of-real-tasks-measured-gains) · [Project Website](https://lh-harness.pages.dev) · [简体中文](README.zh-CN.md)
@@ -30,12 +30,13 @@
 
 > **The model determines what an agent can do in one round. LongHorizon-Harness engineers the loop around it: what to do next, how to verify the result in the real computer, what progress to preserve, and how to continue after failure or context refresh.**
 
-**A Loop Engineering system for Claude Code and Codex. One-command install, ready to run.**
+**A Loop Engineering system for Claude Code, Codex, and DeepSeek Harness. One-command install, ready to run.**
 
 LongHorizon-Harness turns existing agents into long-running computer-use systems. Across desktop apps and the terminal CLI, it continuously recovers the goal and verified state, selects the next bounded step, executes it with a fresh context, checks the actual result, and then checkpoints accepted progress or feeds failure evidence into the next round. It does not train a new model or replace an existing agent; it provides the durable execution loop around one.
 
 ## ✨ News
 
+- **[v0.1.5 · 2026-08-14]** Added phase-1 [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness) CLI support. LongHorizon-Harness can now run `dsh --profile headless` as `--agent deepseek_harness`, with an isolated `DSH_HOME`, role-scoped read/write permissions, DeepSeek API endpoint overrides, normalized JSONL results, and CLI/config/doctor integration. The Web workbench can select DeepSeek Harness and its model independently for each role. GUI computer-use and MCP support will follow in a later phase; see [the CLI setup](#5-or-run-a-task-from-the-command-line).
 - **[v0.1.4 · 2026-08-11]** The new Dashboard has landed: a React/FastAPI workbench you can drive entirely from the browser. Start a task, choose a backend and model per role, answer approvals, send an instruction mid-run, and stop or restart a run. Launch it with `lh-harness web`; see [Run a task in the browser](#4-run-a-task-in-the-browser-recommended).
 - **[2026-08-10]** Added the Terminal-Bench 2.1 evaluation.
 - **[v0.1.3 · 2026-08-07]** Every run now ends with a plain-language reply that answers your task from the verified state alone. Tasks act on the directory you launched from by default, and the console reports each round as it happens.
@@ -102,7 +103,7 @@ LongHorizon-Harness is not tied to a specific model or agent backend. Existing m
 | | Layer | Supported choices |
 |---|---|---|
 | 🧠 | **Models** | Claude, GPT, Qwen, and other models exposed by an agent backend |
-| 🤖 | **Agent backends** | Claude Code, Codex CLI, and custom `AgentAdapter` implementations |
+| 🤖 | **Agent backends** | Claude Code, Codex CLI, DeepSeek Harness (`dsh`, CLI-only in phase 1), and custom `AgentAdapter` implementations |
 | 🎛️ | **Role assignment** | The Manager, Executor, and Auditor can each use a different model or backend |
 | 🖥️ | **Execution environments** | Local, with a pluggable `Environment` protocol |
 
@@ -186,8 +187,8 @@ Steps 1–2 are once per machine; step 3 is once per project. Then run tasks fro
 |---|---|
 | [uv](https://docs.astral.sh/uv/getting-started/installation/) | The recommended isolated install. Skip it if you prefer pip. |
 | Python 3.10 or later | Running the harness. `uv tool install` brings its own; a pip install uses yours. |
-| One agent runtime on `PATH`: [`codex`](https://github.com/openai/codex#installing-and-running-codex-cli) or [`claude`](https://docs.anthropic.com/en/docs/claude-code/getting-started) | Actually executing the work. Install both if you want to mix them across roles. |
-| [Node.js](https://nodejs.org) 20 or later | Only the npm-distributed computer-use plugins. Not needed for `codex-computer-use` or CLI-only tasks. |
+| One agent runtime on `PATH`: [`codex`](https://github.com/openai/codex#installing-and-running-codex-cli), [`claude`](https://docs.anthropic.com/en/docs/claude-code/getting-started), or [`dsh`](https://github.com/deepseek-ai/deepseek-harness) | Actually executing the work. Install more than one if you want to mix backends across roles. |
+| [Node.js](https://nodejs.org) 20 or later | The npm-distributed computer-use plugins. DeepSeek Harness itself currently requires Node.js `^22.19.0` or `>=24.0.0`. |
 
 > **Platform status:** Currently tested on macOS. Windows support is included but has not yet been thoroughly tested.
 
@@ -245,6 +246,48 @@ lh-harness run --task "${TASK}" --agent codex
 
 Explicit CLI arguments such as `--agent` override the matching values in `./.lh-harness/config.toml` for that run; drop them to use the configured defaults.
 
+To use the phase-1 DeepSeek Harness CLI backend, install its official npm package, provide a DeepSeek API key, and select `deepseek_harness`:
+
+```bash
+npm install -g @deepseek-ai/dsh
+# If your npm mirror has not synced the package:
+# npm install -g @deepseek-ai/dsh --registry=https://registry.npmjs.org
+
+dsh --version
+export DEEPSEEK_API_KEY="sk-..."
+# Optional for a private or compatible endpoint:
+# export DEEPSEEK_BASE_URL="https://your-endpoint.example.com"
+
+lh-harness doctor
+lh-harness run --task @task.md --agent deepseek_harness \
+  --model deepseek-v4-flash --no-dashboard
+```
+
+To make DeepSeek Harness the project default, put this in `./.lh-harness/config.toml`:
+
+```toml
+[run]
+agent = "deepseek_harness"
+model = "deepseek-v4-flash"
+dashboard = false
+```
+
+Then use LongHorizon-Harness as usual:
+
+```bash
+lh-harness run --task @task.md
+```
+
+The LongHorizon Web workbench also exposes **DeepSeek Harness (CLI)** in each role's Harness selector and offers `deepseek-v4-flash` plus a custom model ID. Export the provider environment variables before starting the Web server so its worker processes inherit them:
+
+```bash
+export DEEPSEEK_API_KEY="sk-..."
+# export DEEPSEEK_BASE_URL="https://your-endpoint.example.com"
+lh-harness web --workspace-root .
+```
+
+The adapter runs `dsh --profile headless`, gives every run an isolated `DSH_HOME`, uses `workspace-write` for executors, and uses `read-only` for the Manager and auditors. `--api-key` maps to `DEEPSEEK_API_KEY`, `--base-url` maps to `DEEPSEEK_BASE_URL`, and `LH_HARNESS_DSH_BINARY` can select a non-`PATH` binary. DeepSeek Harness is still a developer preview; this phase intentionally does not expose its Web UI, computer-use plugins, MCP config, or `--mcp-add-dir`. Its headless profile currently returns only the final answer, so intermediate DeepSeek tool events are not streamed into the trajectory; the upstream positional task interface also means the task text is visible in the child process argument list while an episode is running.
+
 The agents work in the directory you launched from, so the task acts on your real project. Set `workspace` or `--workspace` to point somewhere else. `./.lh-harness/` itself stays off limits, so the run's own logs and state are never mistaken for task content.
 
 The Dashboard opens in your browser automatically, and the console prints one line per role as the run progresses. At the end you get a plain-language reply that answers your request from the verified state alone, and says so plainly if the task did not finish.
@@ -281,7 +324,7 @@ Task text, run IDs, and API keys are deliberately **not** configurable here; the
 
 | Field | Default | Description |
 |---|---|---|
-| `agent` | `"codex"` | Backend for every role unless a role overrides it: `codex` or `claude_code`. |
+| `agent` | `"codex"` | Backend for every role unless a role overrides it: `codex`, `claude_code`, or `deepseek_harness`. |
 | `model` | `"gpt-5.6-sol"` | Model for every role unless a role overrides it. Must be a model the chosen backend exposes. |
 | `env` | `"local"` | Execution environment. Only `local` today. |
 | `runs_root` | `"./.lh-harness/runs"` | Where run directories are created. Each run gets `<runs_root>/<run-id>/`. |
@@ -443,7 +486,7 @@ lh-harness web --workspace-root .               # Serve the workbench for anothe
 | Option | Description |
 |---|---|
 | `--task` | Task text or `@task.md` |
-| `--agent` | `claude_code` or `codex` |
+| `--agent` | `claude_code`, `codex`, or `deepseek_harness` (CLI-only in phase 1) |
 | `--env` | `local` |
 | `--max-rounds` | Maximum number of Manage-Execute-Audit rounds; the CLI default is 30 |
 | `--dashboard` | Start live monitoring and human intervention |

diff --git a/README.zh-CN.md b/README.zh-CN.md
@@ -4,7 +4,7 @@
 
 ### 面向 Computer-Use Agent 的 Loop Engineering
 
-**只需向 Claude Code 或 Codex 给出一次目标，即可让它跨桌面 App 与终端持续工作数十个小时。**
+**只需向 Claude Code、Codex 或 DeepSeek Harness 给出一次目标，即可让它跨桌面 App 与终端持续工作数十个小时。**
 
 **规划 → 执行 → 验证 → 保存或恢复 → 重复，直到任务真正完成。**
 
@@ -18,7 +18,7 @@
 </p>
 
 [![Python](https://img.shields.io/badge/python-≥3.10-blue?logo=python&logoColor=white)](https://www.python.org/)
-[![Agents](https://img.shields.io/badge/backends-Claude%20Code%20|%20Codex-8A2BE2)](#任意模型任意-agent-后端)
+[![Agents](https://img.shields.io/badge/backends-Claude%20Code%20|%20Codex%20|%20DeepSeek-8A2BE2)](#任意模型任意-agent-后端)
 [![Benchmarks](https://img.shields.io/badge/benchmarks-WeaveBench%20|%20OSWorld%202.0%20|%20Terminal--Bench%202.1-orange)](#数百个真实任务规模化验证)
 
 [Usage](#一条命令全程可见) · [The Loop](#面向真实电脑环境的-loop-engineering) · [Computer Use](#桌面-app-与命令行一个连续任务) · [Results](#数百个真实任务规模化验证) · [Project Website](https://lh-harness.pages.dev) · [English](README.md)
@@ -30,12 +30,13 @@
 
 > **模型决定 Agent 一轮能做什么。LongHorizon-Harness 负责工程化模型外部的执行闭环：下一步做什么、如何在真实电脑中验证、哪些进度可以保存，以及在失败或上下文刷新后如何继续。**
 
-**一套面向 Claude Code 和 Codex 的 Loop Engineering 系统。一条命令安装，开箱即用。**
+**一套面向 Claude Code、Codex 和 DeepSeek Harness 的 Loop Engineering 系统。一条命令安装，开箱即用。**
 
 LongHorizon-Harness 将现有 Agent 变成可长期运行的 computer-use 系统。它在桌面 App 与终端 CLI 之间持续恢复目标和已验证状态、选择下一项有明确边界的工作、用全新上下文执行、检查真实结果，再保存通过验收的进度，或把失败证据带入下一轮。它不训练新模型，也不替换现有 Agent，而是为现有 Agent 提供一套可持续运行的执行闭环。
 
 ## ✨ News
 
+- **[v0.1.5 · 2026-08-14]** 第一阶段已适配 [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness) CLI。LongHorizon-Harness 现在可以通过 `--agent deepseek_harness` 调用 `dsh --profile headless`，并提供隔离的 `DSH_HOME`、按角色划分的读写权限、DeepSeek API 端点覆盖、标准化 JSONL 结果以及 CLI/config/doctor 接入。Web 工作台支持为每个角色分别选择 DeepSeek Harness 及其模型；GUI computer-use 和 MCP 支持将在后续阶段补充。使用方式见 [CLI 配置说明](#5-也可以用命令行运行任务)。
 - **[v0.1.4 · 2026-08-11]** 新版 Dashboard 已上线：基于 React/FastAPI 的工作台，全部操作都能在浏览器里完成——发起任务、为每个角色分别选择后端和模型、处理审批、运行中追加指令、停止或重启任务。用 `lh-harness web` 启动，见[在网页上运行任务](#4-在网页上运行任务推荐)。
 - **[2026-08-10]** 补充了 Terminal-Bench 2.1 评测。
 - **[v0.1.3 · 2026-08-07]** 每次运行结束都会输出一份自然语言回复，只依据已验证状态回答你的任务；任务默认作用于你启动命令的目录，控制台也会实时打印每一轮进展。
@@ -103,7 +104,7 @@ LongHorizon-Harness 不绑定特定模型或 Agent 后端。现有模型和 Agen
 | | 层级 | 支持选项 |
 |---|---|---|
 | 🧠 | **模型** | Claude、GPT、Qwen，以及 Agent 后端提供的其他模型 |
-| 🤖 | **Agent 后端** | Claude Code、Codex CLI，以及自定义 `AgentAdapter` 实现 |
+| 🤖 | **Agent 后端** | Claude Code、Codex CLI、DeepSeek Harness（`dsh`，第一阶段仅 CLI），以及自定义 `AgentAdapter` 实现 |
 | 🎛️ | **角色分配** | Manager、Executor 和 Auditor 可以分别使用不同模型或后端 |
 | 🖥️ | **执行环境** | 本地，并提供可扩展的 `Environment` 协议 |
 
@@ -187,8 +188,8 @@ LongHorizon-Harness 不只展示了几个精心挑选的成功案例。
 |---|---|
 | [uv](https://docs.astral.sh/uv/getting-started/installation/) | 推荐的隔离安装方式。习惯用 pip 可以不装。 |
 | Python 3.10 或更高版本 | 运行 Harness。`uv tool install` 自带 Python；用 pip 安装则使用你当前的。 |
-| `PATH` 上有一个 Agent 运行时：[`codex`](https://github.com/openai/codex#installing-and-running-codex-cli) 或 [`claude`](https://docs.anthropic.com/en/docs/claude-code/getting-started) | 真正执行任务。想按角色混用两个后端就都装上。 |
-| [Node.js](https://nodejs.org) 20 或更高版本 | 只有 npm 分发的 computer-use 插件需要。用 `codex-computer-use` 或纯 CLI 任务时不需要。 |
+| `PATH` 上有一个 Agent 运行时：[`codex`](https://github.com/openai/codex#installing-and-running-codex-cli)、[`claude`](https://docs.anthropic.com/en/docs/claude-code/getting-started) 或 [`dsh`](https://github.com/deepseek-ai/deepseek-harness) | 真正执行任务。想按角色混用多个后端就安装多个。 |
+| [Node.js](https://nodejs.org) 20 或更高版本 | npm 分发的 computer-use 插件需要；DeepSeek Harness 本身目前要求 Node.js `^22.19.0` 或 `>=24.0.0`。 |
 
 > **平台状态：** 目前只在 macOS 上完成了测试；Windows 已支持，但尚未经过详细测试。
 
@@ -246,6 +247,48 @@ lh-harness run --task "${TASK}" --agent codex
 
 命令行上显式传入的参数（如 `--agent`）会覆盖 `./.lh-harness/config.toml` 中的对应配置，仅对本次运行生效；不传则沿用配置文件里的默认值。
 
+使用第一阶段的 DeepSeek Harness CLI 后端时，安装官方 npm 包、提供 DeepSeek API Key，并选择 `deepseek_harness`：
+
+```bash
+npm install -g @deepseek-ai/dsh
+# 如果当前 npm 镜像尚未同步该包：
+# npm install -g @deepseek-ai/dsh --registry=https://registry.npmjs.org
+
+dsh --version
+export DEEPSEEK_API_KEY="sk-..."
+# 使用私有或兼容端点时可选：
+# export DEEPSEEK_BASE_URL="https://your-endpoint.example.com"
+
+lh-harness doctor
+lh-harness run --task @task.md --agent deepseek_harness \
+  --model deepseek-v4-flash --no-dashboard
+```
+
+需要将 DeepSeek Harness 设为当前项目的默认后端时，在 `./.lh-harness/config.toml` 中写入：
+
+```toml
+[run]
+agent = "deepseek_harness"
+model = "deepseek-v4-flash"
+dashboard = false
+```
+
+之后就可以继续使用原来的 LongHorizon-Harness 命令：
+
+```bash
+lh-harness run --task @task.md
+```
+
+LongHorizon Web 工作台也会在每个角色的 Harness 下拉框中显示 **DeepSeek Harness (CLI)**，模型可选择 `deepseek-v4-flash` 或填写自定义模型 ID。启动 Web 服务前先导出 provider 环境变量，确保后续 worker 可以继承：
+
+```bash
+export DEEPSEEK_API_KEY="sk-..."
+# export DEEPSEEK_BASE_URL="https://your-endpoint.example.com"
+lh-harness web --workspace-root .
+```
+
+适配器通过 `dsh --profile headless` 运行，每次 run 使用隔离的 `DSH_HOME`；Executor 使用 `workspace-write`，Manager 和 Auditor 使用 `read-only`。`--api-key` 会映射到 `DEEPSEEK_API_KEY`，`--base-url` 会映射到 `DEEPSEEK_BASE_URL`，也可用 `LH_HARNESS_DSH_BINARY` 指定不在 `PATH` 上的二进制。DeepSeek Harness 仍处于 developer preview；第一阶段暂不接入它的 Web UI、computer-use 插件、MCP 配置或 `--mcp-add-dir`。目前 headless profile 只返回最终答案，因此 DeepSeek 的中间 tool event 不会实时进入 trajectory；上游任务接口采用位置参数，episode 运行期间任务文本也会出现在子进程参数列表中。
+
 Agent 直接在你启动命令的那个目录里工作，任务作用于你的真实项目。需要换到别处时设置 `workspace` 或 `--workspace`。`./.lh-harness/` 本身会被排除在外，本次运行自己的日志和状态不会被当成任务内容。
 
 Dashboard 会自动在浏览器中打开，控制台会逐行打印每个角色的进展。结束时会输出一份自然语言回复，只依据已验证状态回答你的原始任务，没跑完也会如实说明。
@@ -282,7 +325,7 @@ lh-harness check-update
 
 | 字段 | 默认值 | 说明 |
 |---|---|---|
-| `agent` | `"codex"` | 所有角色使用的后端（角色可单独覆盖）：`codex` 或 `claude_code`。 |
+| `agent` | `"codex"` | 所有角色使用的后端（角色可单独覆盖）：`codex`、`claude_code` 或 `deepseek_harness`。 |
 | `model` | `"gpt-5.6-sol"` | 所有角色使用的模型（角色可单独覆盖）。必须是所选后端支持的模型。 |
 | `env` | `"local"` | 执行环境，目前只有 `local`。 |
 | `runs_root` | `"./.lh-harness/runs"` | 运行目录的根路径，每次运行生成 `<runs_root>/<run-id>/`。 |
@@ -444,7 +487,7 @@ lh-harness web --workspace-root .               # 为指定目录启动工作台
 | 参数 | 说明 |
 |---|---|
 | `--task` | 任务文本或 `@task.md` |
-| `--agent` | `claude_code` 或 `codex` |
+| `--agent` | `claude_code`、`codex` 或 `deepseek_harness`（第一阶段仅 CLI） |
 | `--env` | `local` |
 | `--max-rounds` | Manage-Execute-Audit 循环的最大轮数；CLI 默认为 30 |
 | `--dashboard` | 启动实时监控和人工介入功能 |