# linxule/lotus-wisdom-mcp [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/linxule/lotus-wisdom-mcp  
**GitHub Stars:** 37  
**npm Downloads (last month):** 846  
**Views:** 2  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/linxule-lotus-wisdom-mcp

## Description
Contemplative problem-solving using the Lotus Sutra's wisdom framework. Multi-perspective reasoning with skillful means, non-dual recognition, and meditation pauses. Available as local stdio or remote Cloudflare Worker.

## Tools
Capabilities this server exposes over MCP:

- **tag**
- **content**
- **begin**
- **stepNumber**
- **totalSteps**
- **nextStepNeeded**
- **true**
- **isMeditation**
- **meditationDuration**
- **previousJourney**
- **journey**
- **FRAMEWORK_RECEIVED**
- **MEDITATION_COMPLETE**
- **WISDOM_READY**

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "lotus-wisdom-mcp": {
    "command": "npx",
    "args": ["-y","lotus-wisdom-mcp"]
  }
}
```

## Documentation

## What linxule/lotus-wisdom-mcp does

linxule/lotus-wisdom-mcp exposes a contemplative problem-solving workflow based on a cognitive framework inspired by the Lotus Sutra. Rather than producing a single response immediately, the workflow gives a model a sequence of processing techniques and asks it to develop understanding across several steps.

The framework groups tags into wisdom domains. Skillful Means covers direct, gradual, sudden, and related approaches. Non-Dual Recognition addresses recognition, transformation, integration, transcendence, and embodiment. Meta-Cognitive tags support examination, reflection, verification, refinement, and completion. Process Flow includes opening, engaging, and expressing; meditation is represented by the `meditate` tag.

## How it works

A model begins by calling `lotuswisdom` with `tag='begin'` and non-empty `content`. That first response supplies the framework for the inquiry and returns a `FRAMEWORK_RECEIVED` status. Subsequent calls carry the current tag, content, step number, estimated total steps, and whether another step is needed.

Each successful response includes a journey string and domain movement information. The optional `previousJourney` value lets a model continue the path when using the stateless hosted Worker. A meditation step can pause for one to ten seconds; an ongoing pause returns `MEDITATION_COMPLETE`, while a completed inquiry returns `WISDOM_READY`. The model is expected to express the final answer in its own voice after completion.

## Setup and configuration

The server can run locally as an stdio package through `npx`, or remotely through the hosted MCP endpoint at `https://lotus-wisdom-mcp.linxule.workers.dev/mcp`. The remote form is suitable for clients that connect to hosted MCP servers, while the local form keeps the process on the developer's machine.

The `LOTUS_DEBUG=true` environment setting enables per-step trace lines from the local stdio process on stderr. These traces help inspect the thinking sequence without placing diagnostic output in the tool response.

## Tools and capabilities

- `lotuswisdom` processes one contemplative step and returns both JSON text and schema-validated `structuredContent`.
- `lotuswisdom_summary` reconstructs a journey summary from `previousJourney`, including its length, domain path, tags, and brief step content.
- The `contemplate` MCP prompt starts a guided, single-question session.
- The `deep-inquiry` MCP prompt scaffolds a longer movement through process, meta-cognitive, non-dual, and meditation domains.
- MCP ext-apps clients can display an interactive pixel halo with one segment illuminated per step. The visualization includes day and night variants, keyboard access, and reduced-motion support.
- Tool responses expose domain labels, current contemplation text, journey tracking, status values, and an output schema for programmatic consumers.

## Limitations and notes

The framework's results depend on how the connected model and host follow the process. The local server keeps journey information in memory; the hosted Worker does not retain session state, so callers should pass `previousJourney` when continuity is needed. Local calls do not modify user files or external business data, while the hosted Worker records usage analytics. Read-only and idempotency declarations are host hints rather than an absolute guarantee of side-effect-free execution.

_Full upstream README: https://allmcps.com/mcp/linxule-lotus-wisdom-mcp/readme_