Token导航 LogoToken导航

Agent Skill · 研究检索

icp-modeler(ICP 建模器)

icp-modeler 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

icp-modeler 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 23,192。

总安装 4,333周安装 177GitHub Stars 公开资料未说明下载 1,402核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.0.2

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:icp-modeler(ICP 建模器)
来源仓库:https://github.com/drivenautoplex1/icp-modeler
安装命令:
openclaw skills install icp-modeler
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 icp-modeler

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

为抵押贷款和房地产产品设计理想客户档案,提供买家角色与广告定位参数。适用于市场研究与精准广告投放场景,支持 Meta 和 Google 平台定向优化。通过关键词或任务需求触发,返回结构化用户画像与定位建议。安装前需确认权限范围、维护状态及是否涉及联网或文件操作。建议结合原始 README 核验具体用法与数据源准确性。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install icp-modeler;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

icp-modeler 怎么安装?

在终端执行 openclaw skills install icp-modeler。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

icp-modeler 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

icp-modeler 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

icp-modeler 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

icp-modeler 用的人多吗?

公开数据是累计安装 4,333 次,近一周安装 177 次,下载 1,402 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

icp-modeler 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
icp-modeler

ICP 建模器 · 本页

4,333—接触敏感数据MIT-0通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
icp-modeler
description
Generate research-backed Ideal Customer Profiles (ICPs) for mortgage and real estate products. Returns full buyer personas, Meta/Google ad targeting parameters, trigger phrases, content tone, and platform routing — no generic demographics, only actionable targeting intelligence.
version
1.0.2
author
drivenautoplex1
price
0
tags
metadata
openclaw
requires
env
anyBins
primaryEnv
ANTHROPIC_API_KEY
emoji
🎯
homepage
https://github.com/drivenautoplex1/openclaw-skills
install
package
anthropic
bins
[]

ICP Modeler Skill

Know exactly who you're talking to before you write a single word of copy. The ICP Modeler generates battle-tested buyer profiles for mortgage and real estate products — with specific ad targeting parameters, trigger phrases, pain points, and platform routing built in.

Free vs Premium

Free tier (no API key needed):

  • --demo — full crypto-mortgage ICP profile, zero API calls, shows complete output format
  • --list — see all 5 available ICPs
  • --product <name> — display any pre-built ICP profile (works offline)
  • --output json — export any ICP as structured JSON for your own workflows

Premium tier (ANTHROPIC_API_KEY):

  • --generate-content "3 facebook posts" — LLM writes content specifically tuned to the ICP's triggers, tone, and platform presence
  • --generate-content "30s video script" — ICP-targeted video script
  • --generate-content "email subject lines" — subject lines optimized for this buyer's psychology
  • Unlimited content generation via Claude Haiku (~$0.001 per call)

The pre-built profiles alone are worth installing — most "targeting" is just age/income. This gives you psychology.

What this skill does

For each product, generates a complete buyer intelligence package:

  1. Full buyer persona — age, income, location, occupation, pain points, dream outcome
  2. Trigger phrases — exact words and phrases this buyer types into Google and says out loud
  3. Content tone guide — how to speak to this buyer without triggering skepticism
  4. Platform routing — where this ICP actually hangs out (X, Reddit, LinkedIn, Facebook groups)
  5. Meta ad targeting — ages, interests, placements, custom audience strategy
  6. Google Ads targeting — keywords, match types, negative keywords, audience layers

Available ICPs

ProductHeadline
crypto-mortgageThe Crypto Holder Who Won't Sell
credit-repairThe Almost-Ready Buyer
va-loanThe Veteran Who Doesn't Know What They Have
realtor-partnerThe Agent Who Needs a Lender They Can Trust
first-time-buyerThe Overwhelmed First-Timer

Aliases work: crypto, va, credit, realtor, first-time

Usage

# See the crypto-mortgage ICP with zero setup
python3 icp_modeler.py --demo

# List all available ICPs
python3 icp_modeler.py --list

# Pull a full ICP profile
python3 icp_modeler.py --product "crypto mortgage"
python3 icp_modeler.py --product va-loan
python3 icp_modeler.py --product first-time-buyer

# Export as JSON (pipe into other tools)
python3 icp_modeler.py --product credit-repair --output json

# Generate ICP-tuned content (requires ANTHROPIC_API_KEY)
python3 icp_modeler.py --product crypto --generate-content "3 facebook posts"
python3 icp_modeler.py --product va-loan --generate-content "30s video script"
python3 icp_modeler.py --product first-time --generate-content "email subject lines"

# Version
python3 icp_modeler.py --version

Example output (crypto-mortgage ICP)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ICP: Crypto-Backed Mortgage (Fannie Mae / Coinbase / Better)
     "The Crypto Holder Who Won't Sell"
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

DEMOGRAPHICS
  age                  30-48
  income               $120K-$400K household
  location             Target metro, key submarkets and growth corridors
  occupation           Software engineer, finance, entrepreneur, executive

PAIN POINTS
  • Doesn't want to sell crypto and trigger $50K-$500K capital gains tax event
  • Doesn't qualify for traditional mortgage because crypto income isn't W2
  • Feels stuck — 'I have the wealth but can't access it for real estate'

TRIGGER PHRASES
  "don't sell your crypto to buy a house"
  "pledge crypto as collateral"
  "no capital gains event"
  "Fannie Mae crypto"

META AD TARGETING
  Ages:      28-50
  Interests: Cryptocurrency, Bitcoin, Ethereum, XRP, DeFi
  Placement: Facebook Feed, Instagram Feed, Instagram Stories
  Income:    Top 25%

Connecting to other skills

Pipe ICP JSON into the content scorer and content calendar:

# Generate ICP → score the content it suggests → build a calendar
python3 icp_modeler.py --product crypto --output json > icp.json
python3 icp_modeler.py --product crypto --generate-content "5 linkedin posts" | \
  python3 ../content-scorer/score_content.py --stdin

Multi-vertical use

The ICP framework applies to:

  • Any mortgage vertical (FHA, VA, USDA, jumbo, crypto-backed, HELOC)
  • Real estate agents targeting specific buyer profiles
  • Credit repair services targeting pre-approval candidates
  • Financial services with segmented buyer journeys
  • Coaching/consulting with defined client personas

Extend by adding your own ICP dict to the ICPS dictionary in icp_modeler.py.

平台分布

OpenClaw

80.68%
按下载量换算1,131

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills