research · pwguler — Claude Code skill: what it does, how to install
The research Claude Code skill traces library and API behavior to primary sources, or investigates why code took its current shape. It activates for lookups and questions about the history of a setting or constant.
MIT 1,000 installs ★ 10 updated 2026-10-16
npx skills add pwguler/skills --skill research -a claude-codeWhat research does
The skill checks claims about a library, service, or standard against sources maintained by its owner. For questions about code history, it follows commits, pull requests, tickets, documents, and messages tied to the decision. Each claim in the resulting note gets a citation.
It writes a Markdown note to the requested path. If none is given, it uses a location where similar notes live, or picks one and reports it. Failed searches are recorded as gaps. Reading goes to a subagent when requested, when a draft spec triggers AFK decisions, or when the material is over about 2,000 lines or includes images.
What it helps with
- verify API behavior against its owner’s documentation
- trace a code setting to the decision behind it
- write research notes with a source for every claim
- delegate reading of a large specification
Example prompts for the agent
- Check how this API behaves and cite a primary source for every claim.
- Trace the Git history behind this constant and save a Markdown note at
docs/research.md. - Give a subagent this specification to read, then return a note with citations.
How to install research
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill research -a claude-code2Manually: put the files in place
~/.claude/skills/research/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/research/
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill research -a codex2Manually: put the files in place
.agents/skills/research/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill research -a cursor2Manually: put the files in place
.cursor/skills/research/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install pwguler/skills skills/research2Manually: put the files in place
.github/skills/research/SKILL.md1Skills CLI (Vercel)
npx skills add pwguler/skills --skill research -a gemini-cli2Platform CLI
gemini skills install https://github.com/pwguler/skills --path skills/researchManually: put the files in place
.devin/skills/research/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add pwguler/skills --skill research -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about research
How do I install research in Claude Code?
Run npx skills add pwguler/skills --skill research -a claude-code. The skill file is at ~/.claude/skills/research/SKILL.md.
What does research do?
It checks library and API behavior against primary sources, or investigates the history behind code decisions. It returns a Markdown note with citations attached to each claim.
How does research choose sources?
For behavior, it uses vendor documentation, source code, specifications, or a live endpoint. For decision history, it looks at commits, pull requests, tickets, documents, and messages.
Source text
SKILL.md · MIT
--- name: research description: Find out what is true about a library, service, or standard from the owner of the facts. Or find out from its history why code has its shape. Write the answer down with a citation for each claim. Read this skill when the user asks why code, a constant, or a setting has its current form. Use it when the user asks to look something up, or to check how an API or tool really behaves. Use it also to give reading to a subagent while other work continues. --- Do research in this session. A subagent reads instead in three cases. The user asks to give the reading to a subagent. A draft spec launches its AFK decisions. Or the payload is too large for the session: over about 2,000 lines, or images. [Guard the context window](references/guard-the-context-window.md) tells how. The subagent returns the note, not the payload. 1. Sort the question. A question about how a library, service, or standard behaves goes to its owner, in step 2. A question about why our code has its shape goes to its history, in step 3. 2. Read the owner of the fact: the vendor documentation, the source code, the specification, or a live endpoint response. A blog post, a forum answer, or a summary only points you to the owner. It is never the citation. 3. Read the history as [why.md](references/why.md) says. First pin the code. Then read each source that why.md lists. Name every source that the session cannot reach. 4. Return one Markdown note, with the source after each claim. Write the note to the path that the caller asked for. When the caller gave no path, use the place where the repo keeps similar notes. When there is no such place, choose a location, and name it in the reply. Rules: - A citation points at the owner. For behavior, that is the vendor, the source, the spec, or the live endpoint. For intent, it is the commit, PR, ticket, doc, or message that holds the decision. - Never invent a citation. A search that finds nothing is an answer. Write it in the note as a gap.
© pwguler · MIT · pwguler/skills
Similar skills
All skills for Agent setup: memory, context, orchestration →research by other authors comes first.