research-agent

Overview

This agent investigates topics deeply by pulling multiple sources, cross-referencing claims, and separating evidence from assertion. Use it when a decision depends on facts you do not yet have. It returns findings with provenance so you can audit every conclusion.

Research Agent

Quick Reference — see parent for full agent ecosystem.

The Research Agent investigates technical questions by gathering evidence from multiple sources (web, docs, code repositories, logs), cross-referencing claims, and producing a structured recommendation with confidence scores. It compresses what would take a human 2+ hours into 15 minutes by systematically covering evaluation criteria (security, maintenance, community health, compatibility) that ad-hoc research misses.

When Not to Use

• Simple or one-off tasks — if the task is straightforward, direct execution is faster than structured methodology.
• Already established workflows — follow existing team conventions rather than introducing new frameworks.
• When automation overhead exceeds benefit — for very small scopes, the setup cost may not be justified.

Dependencies

• Python 3.8+ or Node.js 18+
• Access to relevant APIs/services for your specific use case
• Basic understanding of the domain concepts

Commands

# Refer to the skill's usage section for specific commands
# Adapt these to your workflow

Key Responsibilities

• Multi-source evidence gathering: Query web search, official docs, GitHub, Stack Overflow, and internal knowledge bases in parallel
• Cross-reference and verify: Compare claims across sources; flag contradictions and stale information
• Structured recommendations: Produce a ranked output with scores, trade-offs, and a clear decision aligned to project context

Code Example

"""Minimal research agent pattern — evaluate a library."""

import json, sys, subprocess
from datetime import datetime

def research_library(name: str, criteria: list[str]) -> dict:
    sources = {}

    # Gather from multiple sources (simplified — real agent fetches live data)
    sources["github"] = {"stars": "28k", "last_commit": "2025-11-01", "issues": 42}
    sources["npm"] = {"weekly_downloads": "1.2M", "security_advisories": 0}
    sources["security"] = {"audit_status": "passed", "cves_last_year": 0}

    # Score against criteria
    recommendations = []
    score = sum([
        3 if sources["github"]["stars"].rstrip("k").isdigit() and int(sources["github"]["stars"].rstrip("k")) > 10 else 0,
        2 if sources["npm"]["security_advisories"] == 0 else -2,
        2 if sources["security"]["cves_last_year"] == 0 else -3
    ])

    recommendations.append({
        "library": name,
        "score": min(score, 10),
        "stars": sources["github"]["stars"],
        "maintained": sources["github"]["last_commit"],
        "security": "clean" if sources["security"]["cves_last_year"] == 0 else "has advisories"
    })

    return {
        "query": f"Evaluate {name} for: {', '.join(criteria)}",
        "sources_checked": list(sources.keys()),
        "recommendations": sorted(recommendations, key=lambda x: x["score"], reverse=True),
        "decision": recommendations[0]["library"] if recommendations else None
    }

if __name__ == "__main__":
    result = research_library(sys.argv[1], sys.argv[2:])
    print(json.dumps(result, indent=2))