agent-researcher

Answers questions by producing a structured research brief with sources and confidence rating.

1|Updated Apr 28, 2026

Search Tags:#planning #documentation #research #codebase #confidence #sources #brain-mcp

Ready to run

npx skills add https://github.com/csackrider/AI-DLC --skill agent-researcher-csackrider

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-researcher
Source: https://github.com/csackrider/AI-DLC/tree/main/plugins/ai-dlc-skills/skills/agents/agent-researcher
Command: npx skills add https://github.com/csackrider/AI-DLC --skill agent-researcher-csackrider

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps teams quickly gather evidence about questions by exploring a codebase, documentation set, brain-mcp collection, or external sources and delivering a structured research brief with sources and a confidence rating.

Core Features & Use Cases

Conducts read-only exploration of codebases, docs, and brain-mcp collections to answer a stated question.
Produces a structured findings report with source references and a confidence level suitable for feature spec writing.
Use Case: In the Plan phase, generate an early research brief to inform design decisions and acceptance criteria.

Quick Start

Research the specified project area and return a structured findings report with sources and a confidence rating.

Frequently Asked Questions about agent-researcher

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I generate a research brief from a codebase for feature planning?▼

To generate a research brief from a codebase, you need a tool that reads code and documentation in read-only mode to produce structured findings with sources and confidence ratings. This process surfaces evidence during plan and design phases to inform feature specs and acceptance criteria.

What is a structured findings report for software design phases?▼

A structured findings report for software design is a concise research brief that answers a stated question by synthesizing evidence from codebases and documentation. It includes exported citations and a clear confidence score to guide design decisions and feature spec writing.

Can I use brain-mcp references to research codebase questions?▼

Yes, you can use brain-mcp references to research codebase questions. The tool conducts read-only exploration of codebases, documentation sets, and brain-mcp collections to answer stated questions and deliver a research brief with a confidence rating.

Does the research process modify my existing code or documentation?▼

No, the research process does not modify your existing code or documentation. It applies read-only exploration across codebases, docs, and external sources to extract findings, ensuring your source files remain untouched while generating the report.

How do I get a confidence score for feature spec research?▼

You get a confidence score for feature spec research by running a read-only exploration of your codebase and documentation. The tool analyzes the gathered evidence and exports a clear confidence level alongside source citations within the structured findings report.

What's the best way to gather evidence for feature specs from documentation?▼

The best way to gather evidence for feature specs from documentation is using an automated research tool that scans docs and code to output a structured brief. This approach provides exported citations and a confidence rating, streamlining early design decisions.