AgentSkillsCN

deep-research

利用浏览器代理、Google AI 与 Perplexity，开展多步骤、深度递归的研究。适用于当简单搜索难以满足需求，且需要经过验证的引用、市场情报或学术论文时使用。

SKILL.md
--- frontmatter
name: deep-research
description: >-
  Orchestrates multi-step deeply recursive research using browser agents, Google AI, and Perplexity.
  Use when simple search is insufficient and verified citations, market intelligence, or academic papers are required.
license: Apache-2.0
metadata:
  version: "1.0.0"
  trit: 1
  author: agent-builder
allowed-tools: [Read, Write, Edit, Bash]
compatibility: Requires Python 3.10+ and Bun for scripts

Deep Research Professional

Overview

High-capability research agent for market intelligence, company analysis, and competitive research. Automates data collection, verification, and report generation using a suite of advanced tools.

Integrated Capabilities

This skill works best when combined with specialized tools:

  1. •Google AI Mode: Use for initial broad discovery, library documentation, and finding recent 2026 data.
  2. •Perplexity (Sonar): Use for intelligent answers, deep reasoning, and comparative analysis.
  3. •NotebookLM: Use for deep semantic analysis of uploaded PDF/Doc reports or synthesizing large documentation sets.
  4. •Deep Research Scripts: Use for structured data extraction and formal report generation.

Capabilities & Commands

1. Academic Research (NEW)

Find peer-reviewed papers, methodologies, and datasets.

bash
# Find papers on a topic
bun scripts/academic-researcher.ts --topic "AI Agents" --focus "methodology"

# Solve CAPTCHA if needed (opens browser)
bun scripts/academic-researcher.ts --topic "AI Agents" --show-browser

2. Intelligent Lookups (Perplexity)

Use Perplexity's Sonar models for reasoning and fact-checking.

bash
# General lookup (Auto-selects Pro/Reasoning model)
python scripts/research_lookup.py "Explain the mechanism of CRISPR off-target effects"

# Force Deep Reasoning
python scripts/research_lookup.py "Compare mRNA vs Traditional Vaccines" --force-model reasoning

3. Company Analysis

Deep dive into specific companies for job prep, investment, or partnership.

bash
# Analyze financial health
bun scripts/company-analyzer.ts --company "Marriott" --focus "financial"

4. Market & Web Research

Broad research on markets, trends, or tools.

bash
# General multi-source search
bun scripts/web-researcher.ts --query "hotel market trends 2024" --depth comprehensive

Advanced Gathering Workflow

For complex topics, use this "Triangulation" pattern:

  1. •

    Scout (Google/Perplexity):

    • •Use google-ai-mode-skill or research_lookup.py to get high-level trends.
    • •python scripts/research_lookup.py "[Topic] key trends 2026"
  2. •

    Deep Dive (Deep Research):

    • •Use web-researcher.ts to crawl specific high-value domains.
    • •bun scripts/web-researcher.ts --query "[Specific Aspect]" --depth comprehensive
  3. •

    Synthesize (NotebookLM):

    • •Upload large PDF reports to NotebookLM.
    • •python ../notebooklm-skill/scripts/run.py ask_question.py --question "Extract financial KPIs" --notebook-url "[URL]"

Methodology & Standards

See references/methodology.md for full details on:

  • •Source Quality Criteria (Authority, Timeliness, Objectivity)
  • •Citation Formats & Link Tracking
  • •Ethical Research Guidelines

Available Tools

  • •scripts/research_lookup.py: NEW Perplexity Sonar integration.
  • •scripts/academic-researcher.ts: Academic paper discovery & PDF extraction.
  • •scripts/company-analyzer.ts: Targeted entity analysis.
  • •scripts/web-researcher.ts: General broad crawler/summarizer.
  • •scripts/report-generator.ts: Formatter for gathered data.