multi-source-research

name	multi-source-research
description	Coordinate quick, standard, or extensive research across web, documentation, and code sources, then synthesize findings with confidence labels.

Research Skill

Multi-source parallel research with confidence-based synthesis.

Trigger Phrases

When user asks:

• "research X for me"
• "find out about X"
• "what's the current status of X"
• "gather information on X"
• "deep dive into X"

Quick Reference

Mode	Speed	Depth	Use Case
quick	~30s	Shallow	Fact check, simple lookup
standard	~2min	Moderate	Typical research
extensive	~8min	Deep	Comprehensive analysis

Command

/ai-dev-kit:research [mode] [query]

Mode Selection

Need quick answer? ─────────────► quick
  │
  ├─ Need current info? ────────► standard
  │
  └─ Need comprehensive? ───────► extensive

Research Types

Type	Sources	Best For
Web	WebSearch	Current events, latest docs
Docs	Local ai-docs/	Library usage, patterns
Code	Codebase grep	Implementation examples

Confidence Levels

Level	Meaning	Source Count
HIGH	Very reliable	3+ agree
MEDIUM	Likely accurate	2 agree
LOW	Needs verification	1 only
CONFLICTING	Check manually	Sources disagree

Cookbook

• cookbook/quick-mode.md - Fast fact-checking pattern
• cookbook/standard-mode.md - Balanced research pattern
• cookbook/extensive-mode.md - Deep research pattern

Reference

• reference/researcher-types.md - Available researcher agents
• reference/synthesis-patterns.md - How findings are combined

Common Gotchas

• Timeout: Extensive mode may timeout on slow connections. Results are still returned.
• Docs only: If web is down, doc-only results are returned with note.
• Conflicting info: Review "Conflicting Information" section when present.