Doc Miner

核心功能：支持摘要生成、内容问答、结构化提取（实体、日期、数字）三种模式，可自动识别任务或由用户指定。典型场景：长文档总结、文档内容答疑、实体提取、纯文本分析、科研文献回顾。输入支持网页/PDF链接或直接粘贴文本，输出对应模式的结构化结果，文档仅临时处理不存储，需通过AIPROX_SPEND_TOKEN认证调用。

unixlamadev-spec

热度 69.57

最近更新 2026-05-11 17:30:11

分类：效率依赖：第三方 API / 内部调用大模型 / 联网浏览搜索应用场景：摘要总结 / 数据处理 / 研究

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add doc-miner

Metadata

name: doc-miner
description: Extract summaries, answers, or structured data from any URL, PDF, or raw text. Auto-detects mode from task.
acceptLicenseTerms: true
metadata: {"clawdbot":{"emoji":"📄","homepage":"https://aiprox.dev","requires":{"env":["AIPROX_SPEND_TOKEN"]}}}

Doc Miner

Extract insights, answers, and structured data from PDFs, webpages, or raw text. Auto-detects the right mode from your task: summarization, Q&A, or structured extraction of entities, dates, and numbers.

When to Use

Summarizing long PDFs or articles
Answering questions about document contents
Extracting named entities, dates, or figures
Analyzing raw text without a URL
Research and literature review

Usage Flow

Provide a url (PDF or webpage) or paste text directly
Optionally specify a task — asking a question triggers Q&A mode; "extract" triggers extraction mode; default is summarization
AIProx routes to the doc-miner agent
Returns mode-specific fields: summary/key_points/word_count, or answer/context/confidence, or entities/dates/numbers

Security Manifest

Permission	Scope	Reason
Network	aiprox.dev	API calls to orchestration endpoint
Env Read	AIPROX_SPEND_TOKEN	Authentication for paid API

Make Request

curl -X POST https://aiprox.dev/api/orchestrate \
  -H "Content-Type: application/json" \
  -H "X-Spend-Token: $AIPROX_SPEND_TOKEN" \
  -d '{
    "task": "extract all dates and key entities",
    "text": "On January 15, 2024, Acme Corp announced a merger with GlobalTech valued at $2.4 billion..."
  }'

Response (extraction mode)

{
  "mode": "extraction",
  "key_points": ["Acme Corp merging with GlobalTech", "Deal valued at $2.4 billion"],
  "entities": ["Acme Corp", "GlobalTech"],
  "dates": ["January 15, 2024"],
  "numbers": ["$2.4 billion"],
  "source_type": "text"
}

Response (summary mode)

{
  "mode": "summary",
  "summary": "Q3 2024 product analytics report covering user metrics and strategic recommendations.",
  "key_points": ["User engagement up 23%", "Mobile conversion 40% below desktop"],
  "word_count": 1240,
  "source_type": "webpage"
}

Trust Statement

Doc Miner fetches and analyzes document contents via URL or processes provided text. Documents are processed transiently and not stored. Analysis is performed by Claude via LightningProx. Your spend token is used for payment only.