Neurodock

MCP办公协作

提供方 tlennon-ie · 包名 @neurodock/cli

办公协作 stdio · 本地进程 OAuth 授权 社区 / 第三方 Claude DesktopClaudeCursorCodex
查看接入方式

A local-first cognitive substrate for neurodivergent professionals. Gives Claude memory, a sense of time, a translator for corporate ambiguity, and a guardrail that refuses to amplify rumination, hyperfocus, or sycophancy. MCP-native. No telemetry. AGPL-3.0-or-later. Self-ID suff

一句话选型结论

按资料字段整理

Neurodock 属于办公协作类 MCP 服务器,stdio · 本地进程,OAuth 授权,可直接接入 Claude Desktop、Claude、Cursor 等,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类办公协作 传输方式stdio · 本地进程 鉴权OAuth 授权 部署方式本地部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

办公协作
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

communication

安装与配置

stdio · 本地进程 OAuth 授权
客户端在本机拉起子进程,数据不出本机
npx --yes @neurodock/cli@latest install-all

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 Cross-platform,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
日程与消息

查日程、发消息、约会议。

表单与审批

填表单、走审批流。

团队协作

同步任务、汇总进展。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 Claude Desktop、Claude、Cursor,换客户端不用重新研究配置。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 使用前要准备好OAuth 授权对应的凭据,并按最小权限配置。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在办公协作场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Neurodock当前 未标注 stdio · 本地进程 本地部署 有 未标注
CowAgent 未标注 stdio · 本地进程 本地部署 有 2026-04-11
Karakeep MCP 未标注 未标注 未标注 无 2026-05-17
agentscope 未标注 stdio · 本地进程 本地部署 有 2026-04-11
note-gen 未标注 stdio · 本地进程 本地部署 有 2026-04-11
Everywhere 未标注 stdio · 本地进程 本地部署 有 2026-04-12

服务器档案

办公协作
字段来自项目仓库与公开索引,中文说明由本站标注
提供方tlennon-ie上游标注的维护方 tlennon-ie
分类办公协作本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式OAuth 授权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时@neurodock/cli · Cross-platform本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验未标注上游最后核验时间

详细介绍

16.8 KB
来自项目仓库与上游目录的原文,未做删改

NeuroDock

A cognitive substrate that remembers, paces, and refuses. Built with neurodivergent professionals, not for them.

5-second TL;DR

  • What it does: gives Claude a memory of your work, a sense of time, a refuse-rumination guardrail, and a translator for corporate ambiguity.
  • Who it's for: neurodivergent people — self-ID only, no diagnosis required, no gatekeeping.
  • How to install: one command, below.

NeuroDock plugs into Claude Desktop / Claude Code / Cursor (any MCP-aware
client). Local-first by default. No telemetry. AGPL-3.0-or-later.

How it fits together

You chat with Claude exactly like normal. Under the hood, Claude calls
NeuroDock's local MCP servers; they read and write your local store; the
result flows back into Claude's reply.

Everything runs on your laptop. Nothing leaves your machine unless you
explicitly turn on a cloud option.

Install```sh

npx --yes @neurodock/cli@latest install-all


One command. It pip-installs the five MCP servers (plus the `neurodock-evals`
harness), wires Claude Desktop / Claude Code / Cursor, and drops a starter
profile at `~/.neurodock/profile.yaml`.

Then, **in any conversation**, try one of:```
What was I working on yesterday?
Plan my morning.
Decompose this goal into atomic tasks.

Claude calls the MCP tools under the hood; you just talk.

> ### 🛑 First-time gotcha: full-quit Claude before testing
>
> Claude only reads its MCP config at startup. Closing the window is not
> enough.
This is the #1 silent failure people hit.
>
> - macOS: Cmd + Q (or Claude menu → Quit Claude).
> - Windows: right-click the Claude icon in the system tray → Quit.
> - Linux: kill the process (pkill -f Claude or quit from the tray).
>
> Then reopen Claude and try one of the prompts above.

Requires Python 3.11+ and Node 22+. Works on macOS, Linux, Windows.

> Why @latest? If you previously ran npm install -g @neurodock/cli,
> npx will silently reuse that older global install instead of fetching the
> current version — and old versions don't have install-all. The @latest
> tag forces npx to resolve against the npm registry every time. Same reason
> we use it on update.

Want to install it step-by-step instead?

NeuroDock ships across two registries: the Python MCP servers (MCP =
Model Context Protocol — the standard that lets Claude call local tools)
on PyPI, and the user-facing CLI on npm. The CLI wraps everything else.

1. Install the six MCP servers```sh
pip install neurodock-mcp-chronometric neurodock-mcp-cognitive-graph neurodock-mcp-task-fractionator neurodock-mcp-translation neurodock-mcp-guardrail neurodock-evals


**2. Wire them into your MCP-aware client**```sh
npx --yes @neurodock/cli@latest init

Detects Claude Desktop, Claude Code, or Cursor and writes the server
entries automatically.

3. Restart Claude (full quit, not close-window — see the callout above).

About the neurodock command — it lives on npm as @neurodock/cli,
not on PyPI. The pip install step gives you the MCP server binaries
that Claude calls over stdio; the CLI is separate. Two ways to run it:```sh

Option A — run via npx, no install

npx --yes @neurodock/cli@latest doctor

Option B — install once, call 'neurodock' from anywhere

npm install -g @neurodock/cli
neurodock doctor


`@neurodock/cli` exposes: `init`, `doctor`, `validate`, `update`, `sync`,
`uninstall`, `host install`, `host uninstall`, `profile show`,
`profile validate`, `install-all`, `install-hooks`, `examples`,
`plugin add/remove/list/enable/disable/validate`.

### Proactive guardrails (optional, recommended)

By default NeuroDock waits for you to ask. That's the wrong shape — an
ND user in hyperfocus is the one least likely to remember to run a
break tool. One command flips it:```sh
neurodock install-hooks --self-test

This wires a small Python script (bundled, stdlib-only — no extra
install needed) into Claude Code's hook system. It then runs silently
on every tool call and auto-fires the chronometric / rumination /
sycophancy heuristics when patterns trip. You get a one-line stderr
banner on the next prompt — never blocked, always dismissible.

For host-agnostic coverage (catches you working in the terminal at
02:00 too), add the standalone daemon:```sh
neurodock install-hooks --install-daemon --self-test


That registers a per-user autostart entry (HKCU Run on Windows,
LaunchAgent on macOS, systemd `--user` on Linux). The daemon polls
every 5 min and surfaces OS-native notifications.

Opt out anytime:```sh
neurodock install-hooks --uninstall          # removes both
export NEURODOCK_GUARDRAILS=off              # disables without removing

The browser extension carries its own equivalent (Phase 2 watchdog) —
it's enabled by default and toggleable from chrome.storage.local via
the popup Settings tab.

See docs/.../proactive-guardrails for the design behind this and the
opt-out matrix.

Want to see it work without installing from PyPI/npm at all?
TESTING_LOCAL.md walks through the from-clone path.

Prefer a plugin, a one-click desktop extension, or the MCP Registry?

The install-all command above is the recommended path. The same servers are
also distributed through three standard channels — all install the identical
local stdio servers, so you can pick whichever fits your client:

Claude Code plugin — bundles the five MCP servers and a set of ND-aware
skills in one step. Requires uv (the servers
run via uvx).```
/plugin marketplace add tlennon-ie/neurodock
/plugin install neurodock@neurodock


**Claude Desktop extension (`.mcpb`)** — one-click install, no config editing.
One bundle per server under `mcpb/`; build with
`.mcpb` onto Claude Desktop → Settings → Extensions. Requires `uv`.

**Official MCP Registry** — each server ships a
registry.modelcontextprotocol.io
under the `io.github.tlennon-ie/*` namespace, so MCP-aware clients can discover
them directly.

guardrail, and `decompose`) are also served over OAuth at a hosted HTTPS
endpoint, so you can use them with no local install and no `uv`:```
https://mcp.neurodock.org/mcp

In Claude (claude.ai or Desktop) → Settings → Connectors → Add custom
connector
→ paste that URL → complete the sign-in prompt. The eight stateless
tools appear immediately. An opt-in memory surface (enable_hosted_storage,
connect_byos_storage, record_fact, recall_entity, …) is also exposed but
does nothing until you explicitly enable storage for your signed-in account. The
personal cognitive graph and neurotype profile are never hosted — they stay
on the local install. Full walkthrough: Hosted server.

The hosted remote server is live at mcp.neurodock.org.

Update```sh

npx --yes @neurodock/cli@latest update


One command. Upgrades all six MCP servers (`pip install --upgrade` /
`uv tool install`), refreshes your wired client configs, and
re-registers the optional native-messaging host. Same flags as
`install-all` (`--client`, `--profile`, `--installer`, `--dry-run`,
`--no-native-host`, `--yes`).

The browser extension auto-updates through the Chrome / Firefox / Edge
store; sideloaded users `git pull` and rebuild.

> Full-quit Claude after updating so it re-reads its MCP config — same
> as first install.

Just want to re-shape client configs without touching package versions?
Use `neurodock sync`.

## See it in action

**The cognitive substrate in a real session** — you talk to Claude normally;
under the hood it calls NeuroDock's MCP tools to remember, pace, and decompose.

https://github.com/user-attachments/assets/bcbf65ce-88dd-4f83-8dc1-d56bdde01bed

**The browser extension** — translating corporate-speak inline on Gmail, Slack,
and the rest, right where you read it, without context-switching back to Claude.

https://github.com/user-attachments/assets/5fa25044-923e-4c25-95bd-f1bd63fb3d6e

> Prefer a written walkthrough? Try the Claude Desktop walkthrough.

## Browser extension (optional)

There's also a browser extension that translates corporate-speak inline on
Gmail, Slack, Linear, Notion, GitHub, Google Docs, and Outlook. It calls
where you'd actually use them — a floating Translate button plus a
right-click menu — so you don't have to context-switch back to Claude
just to decode "let's circle back on this."

You pick the LLM provider. Five options: Ollama (local, default), LM
Studio (local), OpenRouter (including its auto-router), Anthropic, OpenAI.
The API key, if you need one, stays in `chrome.storage.local` and never
leaves the device.

Store submission is still pending (the listing prep is done; the developer
accounts and screenshots aren't). Until then, load it manually:

1. Build it: `pnpm --filter @neurodock/extension-browser run build`
2. Chrome / Edge: go to `chrome://extensions`, turn Developer mode on,
   click **Load unpacked**, and pick `packages/extension-browser/.output/chrome-mv3/`.
3. Firefox: go to `about:debugging`, click **This Firefox** → **Load
   Temporary Add-on**, and pick `manifest.json` inside
   `packages/extension-browser/.output/firefox-mv3/`.

Full per-provider setup walkthrough lives in
`packages/extension-browser/README.md`.

## What's inside

NeuroDock is built around three pillars. Each pillar is made of small,
independent packages that you can use one at a time or all together.```mermaid
flowchart TB
  subgraph cognitive[Cognitive substrate]
    chrono[mcp-chronometric<br/>time + sessions]
    cgraph[mcp-cognitive-graph<br/>memory]
    frac[mcp-task-fractionator<br/>decomposition]
  end
  subgraph comms[Communication layer]
    trans[mcp-translation<br/>tone + meetings]
    ext[extension-browser<br/>Gmail, Slack, ...]
  end
  subgraph safety[Clinical guardrails]
    guard[mcp-guardrail<br/>rumination / hyperfocus / sycophancy]
    clin[clinical<br/>heuristic library]
  end
  skills[skills/<br/>six markdown bundles] --> cognitive
  skills --> comms
  skills --> safety
  cli[cli/<br/>install + manage] --> cognitive
  cli --> comms
  cli --> safety

Prefer the directory tree? Open this.`` neurodock/ ├── packages/ │ ├── mcp-chronometric/ Time + session + break management │ ├── mcp-cognitive-graph/ Persistent memory + entity recall (SQLite) │ ├── mcp-task-fractionator/ Decompose vague goals into atomic tasks │ ├── mcp-translation/ Corporate-speak translator (MCP + browser ext) │ ├── mcp-guardrail/ Rumination / hyperfocus / sycophancy detectors │ ├── skills/ Six SKILL.md bundles activating on phrases │ ├── extension-browser/ WXT-built Chrome / Firefox / Edge extension │ ├── native-host/ Optional native messaging host (extension <-> profile.yaml) │ ├── cli/npx neurodock init` and friends
│ ├── core/ Shared types, profile schema, plugin spec
│ ├── clinical/ Heuristic library for the guardrail server
│ └── evals/ Eval harness + corpus contribution pipeline
├── docs/ Astro Starlight site (deploys to docs.neurodock.org)
├── plugins/ Drop your own plugins here; auto-discovered
└── profiles/ Curated profile presets


## Status

**Public preview shipped.** All three substrate pillars (cognitive,
communication, guardrails) are built, on `main`, and installable from
npm + PyPI. Latest substrate tag: `v0.7.3`.

### MCP servers (PyPI)

- **`neurodock-mcp-chronometric`** — 5 tools, 22 tests, mypy `--strict`.
- **`neurodock-mcp-cognitive-graph`** — 4 tools, SQLite + sqlite-vec
  - fastembed; 4-rung resolution cascade (exact → alias → fuzzy → embedding).
    Latest patch ships friendlier `record_fact` errors so wrong-shape input
    no longer leaks raw Pydantic traces.
- **`neurodock-mcp-task-fractionator`** — 2 tools, 32 tests; ISO 8601
  duration spec.
  baseline plus LLM refinement envelope.
- **`neurodock-mcp-guardrail`** — all three detectors live: rumination,
  hyperfocus, sycophancy (48 tests, public heuristics).
- **`neurodock-evals`** — air-gapped harness, 10 seed corpus examples,
  contribution pipeline.
- **`neurodock-clinical`** — reserved name; importable detector library
  (currently a stub).

### CLI + browser surface (npm)

- **`@neurodock/cli`** — 19 verbs across 7 groups: `init`, `install-all`,
  `examples`, `doctor`, `validate`, `update`, `sync`, `uninstall`,
  `host install/uninstall`, `profile show/validate`, `install-hooks`,
  `plugin add/remove/list/enable/disable/validate`. `install-hooks`
  wires the proactive-guardrail hook into Claude Code and optionally
  registers the standalone daemon.
- **`@neurodock/core`** — profile schema and plugin protocol manifests
  (JSON Schema 2020-12).
- **`@neurodock/native-host`** — optional Chrome Native Messaging host
  for the extension ↔ profile sync.
- **`@neurodock/extension-browser`** — WXT MV3 build for Chrome / Firefox
  / Edge. Seven sites wired. Five real LLM providers: Ollama, LM Studio,
  Anthropic, OpenAI, OpenRouter. Includes the Phase 2 proactive watchdog
  for hyperfocus / late-night / single-host rumination signals.
  **Not yet store-published.**

### Skills + docs

- **Six launch skills** — `adhd-daily-planner`, `audhd-context-recovery`,
  `ocd-decision-finalizer` (beta), `hyperfocus-formatter`, `visual-organizer`,
  `asd-meeting-translator`.
- **Docs site** — builds clean (Astro Starlight), deploys to
  `docs.neurodock.org`.

### Still deferred

- **Browser-store submissions** — Chrome Web Store, Firefox Add-ons, Edge
  Add-ons developer accounts + screenshots; manual step.

## How to actually test it right now

**`TESTING_LOCAL.md`** — step-by-step guide to running this against your
Claude Desktop, today, from a clone. Takes about 5 minutes.

## Documentation

The docs site lives at `docs.neurodock.org`.
Source is at `docs/src/content/docs/`. To preview locally:```bash
pnpm --filter @neurodock/docs run dev
# opens http://localhost:4321

Architecture

The substrate splits into three pillars:

  1. Cognitive substrate — externalises executive function (time,
    memory, decomposition). MCP servers: chronometric, cognitive-graph,
    task-fractionator.
  2. Communication layer — translates corporate ambiguity; rewrites
    outgoing messages for register-appropriate tone; structures meeting
    same prompts in Gmail / Slack / Linear / Notion / GitHub / Docs / Outlook.
  3. Clinical guardrails — detects and intervenes on rumination
    (repeat-validation loops), hyperfocus (escalating session-length
    nudges), and sycophancy (unconditional agreement). MCP server: guardrail.
    Heuristics are public and auditable per ETHICS.md.

All three layers compose via the same MCP protocol the LLM client already
speaks. There's no "NeuroDock app" — the surface is your Claude client.

Design rationale lives in docs/decisions/:

  • ADR 0001 — chronometric tool design
  • ADR 0002 — cognitive-graph tool design
  • ADR 0003 — task-fractionator tool design
  • ADR 0004 — profile schema
  • ADR 0006 — guardrail tool design
  • ADR 0007 — plugin protocol
  • ADR 0008 — distribution & remote strategy

Contributing

CONTRIBUTING.md has the welcome + on-ramp. Pick whichever lane matches
what you want to do:

Smallest first PR (~15 min):

  • Add a test to an existing skill (packages/skills//tests/)
  • Improve a tool's parameter description (the LLM uses these — clearer
    descriptions = better tool use)

One-afternoon PR:

  • Write a new skill — markdown bundle activating on specific phrases.
    Copy any packages/skills// as a template; see
    docs/src/content/docs/contribute/write-a-skill.mdx.
  • Sharpen the task-fractionator heuristics (regex / domain keyword maps).
  • Add a new entity-resolution heuristic (e.g., phonetic) to
    mcp-cognitive-graph.

Multi-day PR:

pack / language pack / theme) per the ADR 0007 plugin protocol. See
docs/src/content/docs/contribute/write-a-plugin.mdx.
Japanese keigo).

No code:

  • Add an anonymised eval example from your own corporate inbox. The
    contribution pipeline lives in packages/evals/ and is the highest-
    leverage non-code contribution.

All PRs run CI: pnpm turbo run lint typecheck test build + uv run pytest

  • uv run mypy --strict packages/.... Every published package has its
    own CHANGELOG. Use Conventional Commits in PR titles (feat:, fix:,
    docs:, test:, chore:, ci:).

Manifesto (short)

  1. Lower friction for users, and for contributors.
  2. Local-first by default; cloud is opt-in.
  3. The user is the authority. Self-ID sufficient.
  4. Composable over monolithic. No god-modules.
  5. Refuse where appropriate. AI that fuels rumination, hyperfocus, or
    anxiety is a regression, not a feature.

Full text in MANIFESTO.md. Ethics framework in ETHICS.md. Governance
in GOVERNANCE.md.

License

AGPL-3.0-or-later. Plugins must declare an AGPL-compatible license
to load — the SPDX whitelist is in ADR 0007.

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
Neurodock 怎么安装?

在客户端按下面的安装命令接入:npx --yes @neurodock/cli@latest install-all。对应的包名是 @neurodock/cli。装完在客户端配置里引用同一个包或端点。

Neurodock 怎么在 Claude Desktop、Claude、Cursor 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Neurodock 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Neurodock 需要 API Key 吗?

资料标注需要OAuth 授权认证,使用前准备好对应凭据,并按最小权限配置。

Neurodock 支持哪些客户端?

资料标注支持 Claude Desktop、Claude、Cursor、Codex、Zed;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Neurodock 能提供哪些能力?

资料显示能力标签包括 communication。具体工具清单以官方文档为准。

Neurodock 需要什么运行环境?

运行时为 Cross-platform;属于本地进程,需要在本机装好运行环境。

Neurodock 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 未标注。