research Skill for Claude Code: Hand Reading Work to a Background Agent
Starts a background agent that reads primary sources for a question and saves the cited findings as one Markdown file in your repo.
- Installs
- 751K
- on skills.sh
- Stars
- 309.2K
- on GitHub
- Files
- 1
- besides SKILL.md
- Updated
- Nov 2026
- last repo commit
research lets you keep working while an agent does the reading. It starts a background agent with a fresh context. That agent checks official docs, source code, specs and first-party APIs instead of write-ups about them, and follows each claim back to its owner.
It writes one Markdown file with a source for every claim. The file goes where your repo already keeps such notes, or somewhere sensible that it names. The SKILL.md names no scripts or keys. The wayfinder skill calls it for research tickets.
Good for
- Developers who need docs or API facts gathered while they keep coding
- Teams that keep cited research notes in the repo
- Anyone planning work that depends on a third-party API
Install research
Claude Code
- 1
Copy this prompt.
Install the research skill in Claude Code. Run this command: npx skills add https://github.com/mattpocock/skills --skill research Then check that it loaded and tell me how to use it.
- 2
Paste it into Claude Code. It shows the change and asks before running it.
Requires Node.js.
From its listing on skills.sh. Source.
What's in the skill
When it loads
Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, docs or API facts gathered, or reading legwork delegated to a background agent.
Key facts
- Author
- Matt Pocock
- Installs on skills.sh
- 750,974 (Nov 2, 2026)
- GitHub stars
- 309,237 (Nov 2, 2026)
- License
- MIT
- Last repo commit
- Nov 1, 2026
- Files next to SKILL.md
- 1 files
- skills.sh audits
- Gen Agent Trust Hub pass, Socket pass, Snyk pass
What it includes
Starts a sub-agent with a fresh context so you keep working
Network ·
Background agentReads official docs, source code, specs and first-party APIs
Network ·
Primary sourcesWrites one Markdown file with the source of every claim
Write ·
Findings fileSaves it where the repo keeps similar notes, or says where it put it
Write ·
Placement
From the official docs
Permissions
Low risk
What it can do
- Read your data
- Create and change things
- Delete things
- Read-only mode
Watch out for
Your question may reach web tools
Keep private names, keys and customer data out of the research question. github.com
It adds a new file to your repo
Check where the file landed before you commit it. github.com
Technical details
| Talks to | Web sources and APIs the agent reads for the question |
|---|---|
| Scopes |
|
Questions
What does research do?
It researches a question in the background and saves cited findings as a Markdown file in your repo. Each claim points to the source that owns it.
How do I install research?
Use the install box at the top. It is a short SKILL.md that does not call other skills.
Does it need an API key?
The SKILL.md names none. The agent reads sources through whatever web and file tools your client gives it.
Where does the file go?
Where your repo already keeps research notes. If there is no convention, it picks a sensible place and tells you.
Can I trust the findings?
The skill asks for primary sources and a citation per claim, which lets you check each one. It is still a model summarising, so open the sources for anything important.
Is it safe?
It reads primary sources and writes one file in your repo, and the SKILL.md describes nothing else. Keep private data out of the question, since it may be passed to web tools your client provides.
Information
updated Nov 2, 2026
- Repository
- mattpocock/skills
- Folder
skills/engineering/research
Sources
Docs, issue trackers and discussions this page is built from.
Show all 4Hide
- research SKILL.md · github.com
- wayfinder SKILL.md · github.com
- research on skills.sh · skills.sh
- mattpocock/skills README · github.com