research

Este extracto es una copia guardada durante la revisión. La fuente externa contiene la versión completa y actual.

---
name: Research
version: 1.5.13
description: "Multi-agent web research with mandatory URL verification, confidence-tagged output, and four depth modes (quick to deep investigation). USE WHEN research, do research, quick research, extensive research, deep investigation, find information, investigate, extract alpha, analyze content, retrieve content, AI trends, enhance content, extract knowledge, web scraping, YouTube extraction, map landscape, competitive analysis, find it, find this, find this product, identify this, what is this, what's that thing, track down, locate, help me find, I can't find X online, can't find it online, source this — never substitute raw WebSearch/WebFetch for a multi-source find/identify/investigate request. NOT FOR people/company/entity deep background, academic papers (use ArXiv), JSON entity extraction, or content-adaptive wisdom extraction (use ExtractWisdom)."
context: fork
background: false
---

## ⚠️ MANDATORY TRIGGER

**When user says "research" (in any form), ALWAYS invoke this skill.**

| User Says | Action |
|-----------|--------|
| "research" / "do research" / "research this" | → Standard mode (3 agents: Claude + Gemini + Perplexity + cross-check) |
| "quick research" / "minor research" | → Quick mode (1 Perplexity agent) |
| "extensive research" / "deep research" | → Extensive mode (7 explorers + 2 verifiers) |
| "deep investigation" / "investigate [topic]" / "map the [X] landscape" | → Deep Investigation (iterative + verification) |

**"Research" alone = Standard mode. No exceptions.**

**Deterministic alternative (EXPERIMENTAL — not yet run in the harness):** `Workflows/research.mjs` ports Standard + Extensive into a Workflow-tool script — fixed researcher roster, single batch URL-verify, cross-checked synthesis. It is parse-verified and contract-checked but has NOT yet had a live harness run, so the prose `StandardResearch.md` / `ExtensiveResearch.md` stay the default path. Do not route real research through the `.mjs` until one smoke run lands. To do that smoke run: `Workflow({ scriptPath: "skills/Research/Workflows/research.mjs", args: { question: "<trivial test>", depth: "standard" } })`. Once it runs clean, drop this experimental caveat.

## Customization

**Before executing, check for user customizations at:**
`~/.claude/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/Research/`

If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.


## 🚨 MANDATORY: Voice Notification (REQUIRED BEFORE ANY ACTION)

**You MUST send this notification BEFORE doing anything else when this skill is invoked.**

1. **Send voice notification**:
   ```bash
   curl -s -X POST http://localhost:31337/notify \
     -H "Content-Type: application/json" \
     -d '{"message": "Running the WORKFLOWNAME workflow in the Research skill to ACTION"}' \
     > /dev/null 2>&1 &
   ```

2. **Output text notification**:
   ```
   Running the **WorkflowName** workflow in the **Research** skill to ACTION...
   ```

**This is not optional. Execute this curl command immediately upon skill invocation.**

# Research Skill

## What It Does

Researches a topic across multiple sources and verifies every claim before delivery. Four depth modes scale from a single fast lookup to a multi-session investigation: Quick (1 agent, ~10-15s), Standard (3 agents cross-checked, ~30-60s), Extensive (7 explorers + 2 independent verifiers, ~60-90s), and Deep Investigation (progressive iteration with a persistent vault, ~3-60min). Output is confidence-tagged: [HIGH] [MED] [LOW] [CONFLICT].

## The Problem

A single AI agent doing research has two failure modes that quietly wreck the result. It hallucinates URLs — confident links that go nowhere, which destroys trust in the whole report. And it answers from one angle, so it parrots whatever the first few search results said and misses conflicts, gaps, and what real pe