Guide · Agent skill
How to use the Deep Research agent skill (deep-research)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The deep-research skill is a set of instructions that tells an AI agent to research a complex question more thoroughly than it would by default. It is part of the ECC project's skills collection, and this guide explains what it does, how to install it, how to ask for it, and where it falls short. By the end you will be able to add the skill to your agent and get a structured answer with a pros and cons table and a list of sources.
Deep Researchaffaan-m/ECC · MIT Download the skill Quick start
- 1Download the ZIP. Download the deep-research skill ZIP from this page.
- 2Unzip the skill. Unzip the file so that SKILL.md sits inside a deep-research folder.
- 3Place the folder. Move the folder to ~/.claude/skills/ for all projects, or to a project's .claude/skills/ folder for one project.
- 4Or upload it in the apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the skill.
- 5Confirm search access. Make sure your agent has Exa search or another web/search tool available.
- 6Ask a research question. Describe the topic, the decision and the sources you want in plain language.
- 7Verify the output. Check the summary, findings and source list against the original sources before relying on them.
What is an agent skill, and what does deep-research make the agent do?
An agent skill is a folder containing a SKILL.md file. The file holds instructions the agent reads when a task matches the skill's purpose. Nothing runs in the background; the agent simply follows the steps written in the file.
The deep-research skill is for complex topics that need several sources. Its description says it performs in-context deep research and produces authoritative answers by curating, cross-referencing and synthesising multiple sources.
- Expands the question into focus areas and angles to investigate
- Scans several kinds of sources: Exa search, GitHub, arXiv papers, blogs and documentation
- Groups notes by theme and looks for common patterns
- Builds a pros and cons table to make trade-offs visible
- Writes a structured answer
Install the deep-research skill
Download the ZIP for the skill from this page and unzip it. The package contains SKILL.md. There are two ways to make the skill available, depending on how you use the agent.
For Claude Code, place the unzipped skill folder in ~/.claude/skills/ to make it available for all your projects, or in a project's .claude/skills/ folder to limit it to that project.
In the Claude apps, you can upload the skill under Settings → Capabilities → Skills instead.
- Personal (all projects): ~/.claude/skills/deep-research/
- Single project: .claude/skills/deep-research/ inside the project folder
- Claude apps: Settings → Capabilities → Skills, then upload the skill
Set up a search tool before you start
The skill's workflow names Exa search and also lists GitHub, arXiv, papers, blogs and documentation as sources. The file does not explain how to set up Exa or any other search access. You need a working web or search tool in your agent environment for the research step to work.
If your agent cannot search the web, the skill can still structure an answer, but it will not be able to gather fresh sources on its own.
- Check that your agent has Exa search or another web/search tool enabled
- Expect the skill to use the sources it can reach, not every source listed
How do you trigger the skill with a natural request?
You do not need a special command. Ask for research in plain language. The skill's usage notes list the kinds of requests it fits: asking for a deep dive on a topic, comparing options or trade-offs, market or competitor research, technical deep dives into frameworks, languages or tools, and questions that need information from several authoritative sources.
Name the topic, the decision you are making and, if relevant, the sources you want the agent to consider.
- "Research the trade-offs between Rust and Go for a CLI tool and list your sources."
- "Do a deep dive on vector databases for a small team. Include a pros and cons table."
- "Compare the top three open-source workflow engines for a competitor overview."
What output should you expect?
The workflow asks the agent to produce a summary of two to three paragraphs, a list of key findings, a pros and cons table, recommendations or best practices, and a list of source references. Use the sources to check the claims before you rely on them.
- Summary (2-3 paragraphs)
- Key findings (bullet list)
- Pros and cons table
- Recommendations or best practices
- Source references
Safety notes: scripts and network access
The skill does not run scripts. It uses network access, because the research step depends on searching and reading sources. Review what the agent searches and reads, especially if your environment holds private material.
The skill itself only gives instructions. It does not guarantee that the output is accurate or current. The agent may summarise inaccurately or miss recent material, so verify important facts yourself.
- Scripts: none included
- Network: yes, used for searching and reading sources
- Verify claims against the listed sources before acting on them
What people use it for
- Developers comparing frameworks, languages or tools before choosing one
- Analysts preparing market or competitor overviews
- Students or researchers who want a first structured survey of a topic with sources listed
- Writing a trade-off analysis for a team decision
Frequently asked questions
How do I install the deep-research skill?
Download the ZIP from this page and unzip it. Place the deep-research folder in ~/.claude/skills/ for all projects, or in a project's .claude/skills/ folder for one project. In the Claude apps, you can instead upload it under Settings → Capabilities → Skills.
Does the deep-research skill need any special setup?
It needs Exa search or another web or search tool in your agent environment, as named in its workflow. The skill file does not explain how to set this up, so check your agent's tool settings.
Is the deep-research skill free?
The source material is licensed under MIT. The file does not mention pricing, so check the project page for any terms beyond the licence.
Does the deep-research skill run scripts?
No. The skill includes only SKILL.md and does not run scripts. It does use network access to search and read sources.
Can I trust the answers from deep-research?
Treat them as a structured first pass. The agent may summarise inaccurately or miss recent material, so check key claims against the sources it lists. For legal, medical or financial decisions, have a qualified professional review the output.
What are alternatives to the deep-research skill?
The source does not compare it with other tools, so this guide cannot make a comparison. If you need only a quick factual answer, a normal question to your agent is faster, since this skill adds extra research steps.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use the deep-research skill when you are comparing options, mapping a topic or preparing an overview and you want a structured answer with sources listed. Skip it for quick facts, and do not rely on it alone for legal, medical or financial decisions or for up-to-date facts without checking the sources yourself.
Download the skill