# RAT (Retrieval Augmented Thinking)

> Use this tool when you need to enhance AI conversation quality and improve response accuracy through structured reasoning and multi-model support. It solves problems such as developing more thoughtful chatbots, enhancing question-answering systems, and facilitating AI-assisted analysis tasks. By providing flexible model selection and persistent conversation context, RAT enables developers and researchers to build more effective AI systems with customizable reasoning visibility.

Canonical page: https://skillsregistry.net/skills/skirano-rat-retrieval-augmented-thinking  
JSON: https://api.skillsregistry.net/v1/skills/skirano-rat-retrieval-augmented-thinking

## Description

This RAT (Retrieval Augmented Thinking) MCP server, developed by Skirano, implements a two-stage reasoning process combining DeepSeek's analysis capabilities with various response models. Built with TypeScript and leveraging the Model Context Protocol SDK, it offers flexible model selection, persistent conversation context, and customizable reasoning visibility. The implementation focuses on enhancing AI responses through structured reasoning, with features like context management and multi-model support. It's particularly useful for developers and researchers working on improving AI conversation quality, enabling use cases such as more thoughtful chatbots, enhanced question-answering systems, and AI-assisted analysis tasks without directly dealing with individual API complexities.

## Trust

- **Trust score (0–1):** 0.95
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** productivity
- **Updated:** 2026-09-28

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/skirano-rat-retrieval-augmented-thinking)
- **Repository:** <https://github.com/newideas99/deepseek-thinking-claude-3.5-sonnet-cline-mcp>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "skirano-rat-retrieval-augmented-thinking"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/skirano-rat-retrieval-augmented-thinking` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/skirano-rat-retrieval-augmented-thinking/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt