Repolode
Submit
MaryCache

MaryCache/imozuru

A Claude Code skill that digs into unfamiliar topics citation by citation until it reaches primary sources, then builds an annotated extract page.

View on GitHub
Stars
652
Forks
27
Open issues
3
Last commit
4 hours ago

What is it?

imozuru is a Claude Code skill for learning an unfamiliar field by following references, links and authors until it reaches the original sources. It borrows ideas from library and information science research on information seeking, such as berrypicking and Ellis's search behaviour model. Instead of summarizing search results, it saves the source texts and records where each needed claim appears in them.

The workflow starts by writing the question and the 3 to 6 claims needed to answer it. Two search agents with different angles collect terms, key people and venues into a map file. The skill then follows citations, cited-by links and authors, saves sources to a local folder, and removes dead or login-gated pages. It stops when every claim has a source or new material stops appearing. A watch mode collects new posts from chosen sites through RSS.

Sources are tagged by type, such as peer-reviewed paper, first-hand account, observation, roundup or official documentation, so readers can judge how much to trust each one. A Python script checks that extracted passages match the saved source lines exactly. Citation chasing uses the OpenAlex API, which needs no key.

Key features

  • 01Maps key terms, people and venues before following citations.
  • 02Saves source texts locally without summarizing them.
  • 03Links each claim to the file and line where its support appears.
  • 04Tags sources by type, such as paper, first-hand account or official document.
  • 05Provides a watch mode that collects new posts from chosen sites via RSS.
  • 06Generates an HTML extract page checked against the saved sources.

Use cases

  • Getting oriented in a field where you do not know the vocabulary or the main researchers.
  • Researching niche topics that live mostly in blogs and videos rather than papers.
  • Building a source-backed reading list with notes on where each claim is supported.
  • Keeping track of new posts from specific communities over time.

Topics

#claude-code#agent-skill#research#literature-review#information-retrieval#citation-tracking#python

FAQ

What is imozuru?
A Claude Code skill that digs into unfamiliar topics citation by citation until it reaches primary sources, then builds an annotated extract page. imozuru is a Claude Code skill for learning an unfamiliar field by following references, links and authors until it reaches the original sources. It borrows ideas from library and information science research on information seeking, such as berrypicking and Ellis's search behaviour model. Instead of summarizing search results, it saves the source texts and records where each needed claim appears in them.
Is imozuru free to use? What license does it have?
imozuru is open source under the MIT license. Check the license terms before using it in commercial products.
Is imozuru still maintained?
Actively maintained. The last commit to imozuru was on Oct 19, 2026.
What can I use imozuru for?
Getting oriented in a field where you do not know the vocabulary or the main researchers. Researching niche topics that live mostly in blogs and videos rather than papers. Building a source-backed reading list with notes on where each claim is supported. Keeping track of new posts from specific communities over time.
Where can I get imozuru?
The source code and installation instructions are on GitHub: https://github.com/MaryCache/imozuru

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects