Hyperresearch skill
A Claude Code skill that turns the agent into a researcher: multi-step source collection, citation verification, and a wiki report that grows session by session
- 4,715 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Accesses external web sources and scientific databases
- Requires a model key and consumes tokens on a multi-step run
Install
- Works in
Manual install
pip install hyperresearch && hyperresearch installInstalls the package and registers the skill in the project. To install for all sessions, add the --global flag.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Accesses external web sources and scientific databases
- Requires a model key and consumes tokens on a multi-step run
This is third-party code. Review the repository files before installing.
What it does
Hyperresearch runs a multi-step research pipeline in Claude Code: it breaks down the query into subpoints, conducts parallel source searches, identifies contradictions, explores disputed areas, critiques and rewrites drafts multiple times, then verifies each citation against the text. Besides the regular web, it searches scientific databases like OpenAlex, Crossref, CORE, ClinicalTrials.gov, and SEC EDGAR, and tries to find legal open copies of paid articles. All found sources are saved in a local markdown and sqlite database, which the next session reuses without downloading them again.
Who it is for. For those preparing extensive research reports with verifiable sources through Claude Code: analysts, content editors, longread authors.
Good fit when
- You need a report on a topic with dozens of verified sources rather than a memory-based summary
- It is important to distinguish independent sources from reprints of the same press release
- There are paid articles and you need to find their legal open copies
Not a fit when
- You need a quick answer to a simple question without source collection
- Python 3.11 or newer is not installed, or there is no time for a multi-step run
Example request
Compile a research report on the impact of remote work on developer productivity with verified sourcesLimitations
The skill is distributed as a package for Python 3.11–3.14 and designed for Claude Code. A full run with critics and citation checks takes significant time and consumes many tokens and external source calls. Some scientific databases may be unavailable or slow without a VPN.
How to disable. Uninstall the package with pip uninstall hyperresearch; for global installation, also remove the added system text from Claude Code settings.
FAQ
Are sources downloaded again in a new session?
No, each found source is stored locally in markdown and sqlite; the next session checks this storage first before accessing the network.
Can it be used outside Claude Code?
The entry point is implemented as a Claude Code skill via its Skill tool; the README describes installation specifically for Claude Code.
Discussion
No questions or reviews yet.
Related
SkillsDeep Research
Deep Research Skill for Claude Code
A deep research skill for Claude Code: multi-phase search, claim verification and cited reports in Markdown, HTML and PDF
A Claude Code skill that turns "research this topic" into a multi-phase investigation: hypotheses, parallel search, a file per source and citation checks
A set of skills for searching scientific databases, PubMed, arXiv, ChEMBL, ClinicalTrials and more, with plain-language queries.
A skill for Claude Code and Codex: one query goes to 52 search channels and the results merge into one deduplicated JSON
Three search skills: four-source search scored by intent, deep thread tracing and clean Markdown extraction from pages.
A CLI and MCP server that sends one question to search and AI sources and saves answers, citations and cost in a run.
README in short
The README presents Hyperresearch as a sixteen-step research pipeline: from query analysis and broad search to a contradiction graph, critic opponents, targeted edits, and final readability check. The separate hpr scholar search command combines multiple scientific databases in one query. Installation is via pip, run with /hyperresearch inside Claude Code; the run can be paused and resumed at the same step thanks to a launch manifest.