home / skills / robdtaylor / personal-ai-infrastructure / research

Research skill

/skills/Research

This skill facilitates multi-source research using parallel queries from Perplexity, Gemini, and Claude to deliver comprehensive findings.

npx playbooks add skill robdtaylor/personal-ai-infrastructure --skill research

Review the files below or copy the command above to add this skill to your agents.

Files (11)
SKILL.md
5.5 KB
---
name: Research
description: Multi-source parallel research using Perplexity, Gemini, and Claude. Three modes - Quick (1 query), Standard (3 queries), Extensive (8 queries). USE WHEN user says 'research', 'investigate', 'find out about', 'do research on', or requests comprehensive information gathering.
---

# Research Skill

## CRITICAL: Use the Research CLI for Multi-Source Research

**For ANY research request, use the `research` CLI via Bash:**

```bash
# All sources in parallel (RECOMMENDED - fastest)
~/.claude/bin/research --all "your research query"

# Individual sources
~/.claude/bin/research --perplexity "query"    # Fast web search
~/.claude/bin/research --gemini "query"        # Multi-perspective
~/.claude/bin/research --claude "query"        # Detailed analysis (slower)
```

**DO NOT use WebSearch directly for research requests - use the CLI above.**

---

## Three Research Modes

### QUICK RESEARCH MODE
- **Trigger:** User says "quick research" or simple queries
- **Method:** Single `research --all "query"` call
- **Time:** ~15-30 seconds
- **Result:** 3 source results (Perplexity, Claude, Gemini)

### STANDARD RESEARCH MODE (Default)
- **Trigger:** Default for most research requests
- **Method:** 3 parallel Bash calls with different query angles
- **Time:** ~30-45 seconds
- **Result:** 9 source results (3 angles × 3 sources)

### EXTENSIVE RESEARCH MODE
- **Trigger:** User says "extensive research" or "deep dive"
- **Method:** 8 parallel Bash calls covering diverse angles
- **Time:** ~45-90 seconds
- **Result:** 24 source results (8 angles × 3 sources)

---

## Standard Research Workflow

### Step 1: Decompose the Question

Break the user's question into 3 focused sub-questions:
- Core facts and recent developments
- Context, implications, and analysis
- Contrarian views, controversies, or edge cases

### Step 2: Launch Parallel Research (ONE MESSAGE)

**CRITICAL: Use a SINGLE message with multiple Bash tool calls for parallel execution**

```bash
# Launch all 3 in parallel (single message with 3 Bash calls)
~/.claude/bin/research --all "sub-question 1 focusing on core facts"
~/.claude/bin/research --all "sub-question 2 focusing on implications"
~/.claude/bin/research --all "sub-question 3 focusing on controversies"
```

### Step 3: Parse JSON Results

Each call returns JSON:
```json
{
  "source": "perplexity|claude|gemini",
  "query": "the query",
  "timestamp": "ISO timestamp",
  "success": true,
  "content": "research findings...",
  "citations": ["url1", "url2"],
  "duration_ms": 1234
}
```

For `--all`, returns array of 3 results.

### Step 4: Synthesize Findings

**Confidence Levels:**
- **HIGH CONFIDENCE**: Corroborated by 2+ sources
- **MEDIUM CONFIDENCE**: Found by one source, seems reliable
- **LOW CONFIDENCE**: Single source, needs verification

**Structure:**
```markdown
## Key Findings

### [Topic Area 1]
**High Confidence:**
- Finding X (Sources: perplexity, claude)

**Medium Confidence:**
- Finding Z (Source: gemini only)

## Source Attribution
- **Perplexity**: [web/current events]
- **Claude**: [detailed analysis]
- **Gemini**: [multi-perspective synthesis]

## Conflicting Information
- [Note any disagreements]
```

---

## Extensive Research Workflow

### Step 1: Generate 8 Diverse Angles

- Core facts and current state
- Historical context and evolution
- Technical deep-dive
- Practical implications and applications
- Contrarian views and criticisms
- Future predictions and trends
- Cross-domain connections
- Edge cases and unusual perspectives

### Step 2: Launch 8 Parallel Queries

```bash
# All 8 in parallel (single message)
~/.claude/bin/research --all "angle 1: core facts about [topic]"
~/.claude/bin/research --all "angle 2: historical context of [topic]"
~/.claude/bin/research --all "angle 3: technical deep-dive on [topic]"
~/.claude/bin/research --all "angle 4: practical implications of [topic]"
~/.claude/bin/research --all "angle 5: criticisms around [topic]"
~/.claude/bin/research --all "angle 6: future predictions for [topic]"
~/.claude/bin/research --all "angle 7: cross-domain connections to [topic]"
~/.claude/bin/research --all "angle 8: edge cases of [topic]"
```

---

## API Keys Required

Located in `~/.claude/.env`:

| Feature | API Key | Required |
|---------|---------|----------|
| Perplexity | `PERPLEXITY_API_KEY` | For --perplexity |
| Gemini | `GOOGLE_API_KEY` | For --gemini |
| Claude | None | Built-in WebSearch |

---

## Critical Rules

### Parallel Execution
- ✅ Launch ALL research queries in ONE message (parallel Bash calls)
- ✅ Each query covers a different angle
- ❌ DON'T launch sequentially (kills speed benefit)
- ❌ DON'T wait between queries

### Error Handling
- If a source fails, proceed with successful results
- Note failures in the final report
- Check `success: false` in JSON for errors

---

## Research Metrics Template

Include at end of every research report:

```markdown
**📈 RESEARCH METRICS:**
- **Mode:** [Quick/Standard/Extensive]
- **Total Queries:** [X] (angles × 3 sources)
- **Sources:** Perplexity, Claude, Gemini
- **Confidence Level:** [High/Medium/Low] ([%])
```

---

## Example: Standard Research

**User:** "Research quantum computing developments"

**Kai executes:**
```bash
# 3 parallel Bash calls in ONE message
~/.claude/bin/research --all "quantum computing recent breakthroughs 2026"
~/.claude/bin/research --all "quantum computing practical applications timeline"
~/.claude/bin/research --all "quantum computing challenges limitations"
```

**Result:** 9 source queries (3 angles × 3 sources) in ~30-45 seconds.

Overview

This skill runs multi-source parallel research using Perplexity, Gemini, and Claude. It offers three modes—Quick (1 query), Standard (3 queries), and Extensive (8 queries)—to match the depth of the request. Use this when you ask to research, investigate, find out about, or do research on a topic. The skill returns synthesized findings, source attributions, confidence assessments, and research metrics.

How this skill works

For any research request the skill triggers the research CLI to run parallel queries across the three sources. Quick mode issues a single multi-source query; Standard runs three angle-focused queries in parallel; Extensive runs eight diverse-angle queries in parallel. Each CLI call returns JSON results (source, query, content, citations, success, timestamp) which the skill parses, corroborates, and synthesizes into a concise report with confidence labels.

When to use it

  • When you say 'research', 'investigate', 'find out about', or ask for a comprehensive information gathering.
  • When you need corroborated facts and up-to-date citations from multiple web-aware models.
  • When you want a quick summary (Quick) versus a balanced multi-angle view (Standard) or a deep dive (Extensive).
  • When you need explicit source attribution and confidence levels for decisions or reporting.

Best practices

  • State the desired mode (quick, standard, extensive) or allow the default Standard for most requests.
  • Break broad topics into a focused question or let the skill decompose into core facts, implications, and contrarian views.
  • Allow the skill to launch all queries in one parallel batch—this ensures speed and consistent result timing.
  • Scan the returned citations and flagged failures before acting on low-confidence findings.
  • Provide follow-up clarifying prompts to refine angles or request deeper technical details.

Example use cases

  • Quickly summarize the latest developments on a breaking tech story using the Quick mode.
  • Produce a balanced 3-angle briefing on market implications of a new regulation with Standard mode.
  • Run an 8-angle deep dive into a complex scientific topic, including historical context and edge cases, using Extensive mode.
  • Gather corroborated sources and confidence levels before drafting a report or presentation.
  • Investigate conflicting claims across sources and produce a concise conflict analysis for stakeholders.

FAQ

Which mode should I pick by default?

Use Standard for most requests; Quick for short factual lookups and Extensive for full deep dives.

What happens if one source fails?

The skill proceeds with successful results, notes the failure in the report, and adjusts confidence accordingly.