Token导航 LogoToken导航TokenDH.com
研究检索需要联网clawhub未标认证来源可访问clear审计通过

mortgage-kb-ragmortgage KB RAG 搜索

Agent Skill

用于搭建或维护带检索增强的 RAG 工作流,适合让 Agent 处理知识库问答、向量检索、来源引用和事实核查。它可以辅助整理数据接入、Embedding、向量库、召回参数和回答生成流程。使用时需要确认数据来源、更新频率、召回阈值和引用展示方式,避免把未命中的资料或过期内容包装成确定事实。

总安装

4,044

周安装

162

GitHub Stars

公开资料未说明

下载量

1,309
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:mortgage-kb-rag(mortgage KB RAG 搜索)
来源仓库:https://github.com/drivenautoplex1/mortgage-kb-rag
安装命令:
openclaw skills install mortgage-kb-rag
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install mortgage-kb-rag

简介

mortgage-kb-rag 基于精选知识库即时回答抵押贷款、信贷与购房相关问题。

  • 适用于客服问答、内部培训资料库或政策合规核查用途。
  • 支持向量检索与来源引用,增强回答可信度。
  • 安装命令:openclaw skills install mortgage-kb-rag;需初始化向量数据库。
  • 应定期更新知识库内容,防止过时政策误导用户。

SKILL.md

name
mortgage-kb-rag
description
Instant answers to mortgage, credit, and home-buying questions from a curated knowledge base. Ask anything about the buying process, credit repair, loan programs, or terminology — get accurate, sourced answers in seconds.
version
1.0.2
author
drivenautoplex1

Mortgage Knowledge Base RAG

Answer any mortgage or home-buying question instantly from a curated expert knowledge base. No hallucination — every answer is grounded in the indexed documents.

What It Does

Given a plain-English question, this skill:

  1. Searches the mortgage and real estate knowledge base using semantic similarity
  2. Retrieves the most relevant passages (with source citations)
  3. Synthesizes a clear, accurate answer — citing which document it came from
  4. Flags gaps — if the question isn't in the knowledge base, says so rather than guessing

Use Cases

  • MLOs answering client questions quickly without Googling
  • Real estate agents prepping for buyer consultations
  • First-time buyers who need straight answers about the process
  • Agents building FAQ content for their websites
  • Verifying what a lender told you is actually correct

Inputs

{
  "query": "What credit score do I need to buy a house?",
  "category": "credit|loans|process|terminology|programs|investing",
  "detail_level": "brief|standard|detailed",
  "audience": "buyer|agent|mlo|investor"
}
  • query (required): Any natural-language question about mortgages, credit, or home buying
  • category (optional): Narrows search to a specific knowledge domain
  • detail_level (optional, default: standard): How much detail to return
  • audience (optional, default: buyer): Tailors language complexity to the reader

Outputs

{
  "answer": "To qualify for a conventional loan, most lenders require a minimum 620 credit score. FHA loans accept scores as low as 580 with 3.5% down, or 500 with 10% down. However, a score above 740 gets you the best rates — typically 0.5-1.0% lower APR than a 620, which translates to $150-300/month on a $350K loan.",
  "confidence": "high",
  "sources": [
    {
      "document": "First-Time Buyer Roadmap",
      "section": "Phase 1: Financial Snapshot",
      "relevance": 0.94
    },
    {
      "document": "90-Day Credit Repair Playbook",
      "section": "Home Price Math: Real Dollar Impact by Credit Tier",
      "relevance": 0.87
    }
  ],
  "related_questions": [
    "How do I improve my credit score in 90 days?",
    "What is the difference between FHA and conventional loans?",
    "How much does a 1-point credit score difference affect my mortgage rate?"
  ],
  "knowledge_gap": false
}

Knowledge Base Coverage

The indexed knowledge base covers:

Credit & Financing

  • Credit score tiers and their rate impact (real dollar math)
  • The 5 credit factors and which to improve first
  • FHA, conventional, VA, USDA loan programs
  • Down payment assistance programs (national + state-level)
  • Debt-to-income ratio calculation and limits
  • What triggers hard vs. soft credit inquiries

The Buying Process

  • Pre-approval vs. pre-qualification (what agents actually care about)
  • Making competitive offers in different market conditions
  • Inspection timelines and what to push back on
  • Closing costs breakdown (who pays what)
  • The 45-day closing timeline step-by-step
  • What NOT to do after going under contract

Credit Repair

  • Dispute letter process (FCRA rights)
  • Pay-for-delete negotiation strategy
  • Goodwill letter templates and when they work
  • Credit utilization optimization (fastest score mover)
  • Authorized user strategy
  • 48-hour rapid rescore process

Market & Terminology

  • 40+ glossary terms (escrow, PMI, LTV, DTI, rate lock, etc.)
  • How to read a Loan Estimate (LE) vs. Closing Disclosure (CD)
  • Rate lock timing strategy
  • Understanding APR vs. interest rate

Lead Nurture & Follow-Up (MLO/Agent Tools)

  • 90-day credit repair client journey
  • Follow-up script patterns by lead stage
  • Objection handling (rates, timing, affordability)

Pricing

TierPriceQueries
Single query$11 question
Pack$710 questions
Unlimited (30 days)$19Unlimited

Technical Notes

  • Knowledge base: ~450 Q&A pairs + sourced document passages
  • Retrieval: TF-IDF + semantic re-ranking
  • Latency: <2s for standard queries
  • Hallucination guard: returns knowledge_gap: true if confidence < 0.70
  • Updates: knowledge base refreshed monthly with new loan program data

适合场景

01

研究助手

02

事实核查

03

知识库问答

04

带来源的搜索总结

能力概览

能力 1

组合搜索和大模型调用

能力 2

支持多来源检索和总结

能力 3

强调引用来源和事实核查

能力 4

适合研究型 Agent 流程

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

89.52%
按下载量换算1,172

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills