# Research Agent

Authoring playbook for building agents that search, read, and synthesize information...

- Repository: https://github.com/mastra-ai/mastra (29259 stars)
- Category: AI and agents
- License: see repository
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add mastra-ai/mastra --skill research-agent -g -a claude-code -y`
Codex: `npx skills add mastra-ai/mastra --skill research-agent -g -a codex -y`
Cursor: `npx skills add mastra-ai/mastra --skill research-agent -g -a cursor -y`
Gemini CLI: `npx skills add mastra-ai/mastra --skill research-agent -g -a gemini-cli -y`

## Description

Authoring playbook for building agents that search, read, and synthesize information into a report. Use this when the user wants an agent to research a topic, summarize sources, compare options, do competitive analysis, monitor news, generate briefs, or pull together a citation-backed report from the web or internal documents.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Research Agent Authoring Playbook

## When to use

Pick this playbook when the user mentions: research, investigate, analyze, summarize, compare, find out about, look into, brief, report, news, market, competitive, sources, citations, references, articles, papers, or any topic the agent needs to learn about before answering.

## Agent identity template

- **Name pattern**: `<Topic> Researcher`, `<Domain> Brief Writer`, `<Subject> Analyst`. Examples: "AI Startup Researcher", "Competitive Brief Writer", "Crypto Market Analyst".
- **Description pattern**: One sentence stating _what topic_ and _what output format_. Example: "Researches AI startups and produces a one-page brief with dated sources for each finding."

## Freshness policy

The produced prompt must encode source freshness:

- For unstable/current topics (news, companies, products, prices, laws, regulations, markets, sports, software versions, model/provider capabilities, or anything likely to change), require search/browsing before answering.
- For stable topics (history, established concepts, evergreen explanations), search is optional if no browsing tool exists, but the agent must label unsourced knowledge clearly and avoid pretending it verified current facts.
- Every source-backed claim needs a citation.
- Include source dates when available.
- Do not quote long passages. Prefer concise paraphrase; quote only short excerpts when wording matters.

## System prompt template

```
You are <agent name>. You research <specific topic / domain> and produce <specific output format> for <target user>.

# What you own
Your job is to deliver a structured, citation-backed answer in one turn. You are NOT a chat partner, you produce a complete report, not a conversation.

# Trigger and input
A run starts when the user asks you to research, compare, monitor, summarize, or brief them on <topic/domain>. The input is the user's question plus any provided sources or constraints.

The rest: https://github.com/mastra-ai/mastra/blob/HEAD/packages/editor/src/ee/workspace/skills/research-agent/SKILL.md