Research Weave skill
Research Weave turns a pile of papers into a reading queue, an evidence map and a sharper research question
- 80 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill reads papers and reaches the network through agent tools and adds no permissions of its own
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/research-weaveDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill reads papers and reaches the network through agent tools and adds no permissions of its own
This is third-party code. Review the repository files before installing.
What it does
The skill gives an agent a repeatable way to answer four questions: which papers deserve a close read, what is established in them, whether the findings agree and what this changes for your research. It traces how the field developed, connects claims across works, weighs converging and conflicting evidence and returns a reading queue with reasons. Five lenses: intellectual history, evidence networks, research relevance, converging support and contradictions. The skill counts supporting papers, shared result families and confirmed independent contributions separately, so repeated reports of one experiment count as one basis. A saved run usually holds `report.md` with reading decisions and the project position, `evidence.json` with claims, provenance and relationships and `search-log.md` with searches, screening and the stopping reason. Scout mode starts with a short report, and a full network is built for Map or Update or when it changes the decision. Install for Codex: copy `skills/research-weave` into `~/.agents/skills/` and call `$research-weave`.
Who it is for. Researchers and graduate students who need to decide which papers to read first and what they change in their project.
Good fit when
- A pile of papers on a topic needs a reading order with reasons and the right section
- Two works disagree and you must see whether they are comparable
- You need to check how many independent contributions back a claim
- Before choosing between a revision loop and resampling you need a position on the project
Not a fit when
- You need a finished literature review as an article: the output is a reading queue and an evidence map
- You need to check online search completeness: the author did not benchmark search recall
- The topic is not research: the lenses are built for scientific papers
Example request
Use $research-weave: decide whether to build a revision loop or spend the same inference budget on resamplingLimitations
- The author states the verification boundary: online search completeness and comparative quality and cost were not measured.
- Behavior is confirmed on small supplied corpora and a bounded historical walkthrough.
- Codex install is described in detail, while other environments need their own skills folder and invocation method.
- Working with live search needs a network tool on the agent side.
- The result helps decide what to read and what to expect from the literature, while conclusions about the truth of the work stay with the researcher.
How to disable. Delete the `~/.agents/skills/research-weave` folder or the same one in the project directory.
FAQ
What results does the skill give?
A reading queue, a research map, an evidence synthesis, a tension analysis and a project position.
What are the five lenses?
Intellectual history, evidence networks, research relevance, converging support and contradictions.
Where are the run files?
A saved run holds `report.md`, `evidence.json` and `search-log.md`: the report, the claim base and the search log.
How do I install it in Codex?
Clone the repository and copy `skills/research-weave` into `~/.agents/skills/` then call `$research-weave`.
Does the skill count repeated reports as separate evidence?
No, repeated reports of one experiment remain one basis, and independent contributions are counted separately.
Discussion
No questions or reviews yet.
Related
SkillsA marketplace of five Claude Code plugins for researchers: literature search, research design, project memory and manuscript work
A library of 280+ research skills: literature search, experiment design, analysis, reference checks and paper preparation
Seven Codex skills carry research from topic choice and paper reading to reviewer replies and an editable presentation.
A skill for master and PhD students that runs research through novelty and reality gates, from review to claim checks
A Claude Code skill for scientific literature reviews: an interview, Scopus and OpenAlex search, DOI checks and a final docx file
A research loop that keeps evidence: a literature review, paper review or survey runs in steps with checks and resume.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Research Weave (this card) | no | 80 | |
| literature-survey-skill | no | 115 | a literature review through NotebookLM and `/survey` modes, while Research Weave weighs independent evidence and contradictions |
| kelip-paper-reading | no | 355 | guides a reader through one paper in six steps, while here the analysis covers a whole field |
README in short
The README gives the output formula and verification boundary, explains five lenses, lists results, shows a walkthrough on three self-correction papers and a fictional source packet for testing. Then come the Codex install and a sample prompt.