Cenrax/researchswarm

An agentic system that reads research papers (local PDFs or arXiv), generates implementation plans, writes code, and reviews it — all orchestrated by a central Director agent built on the Claude Agent SDK.

24Stars on the repository
15Mods indexed here, across every type
4mo agoLast push, which is what freshness is scored on
noneNo LICENSE: all rights reserved, so bodies are not copied

architecture-design

01

Cenrax/researchswarm

Skill Claude CodeCodex

Design the software architecture and component breakdown for implementing research paper techniques. Use this when creating the implementation plan to define system structure.

not rated 24 4mo ago A 25 tokens

dependency-analysis

02

Cenrax/researchswarm

Skill Claude CodeCodex

Analyze and select third-party libraries and frameworks needed to implement the research paper. Use this when planning dependencies to avoid version conflicts and bloat.

not rated 24 4mo ago A 28 tokens

extract-equations

03

Cenrax/researchswarm

Skill Claude CodeCodex

Extract and catalog all mathematical equations, formulas, and pseudocode from a research paper. Use this when you need precise mathematical details for implementation.

not rated 24 4mo ago A 28 tokens

iterate-feedback

04

Cenrax/researchswarm

Skill Claude CodeCodex

Send code back to the coder agent with reviewer feedback for a rework iteration. Use this when the reviewer reports NEEDSCHANGES or FAIL, and the user approves a rework cycle.

not rated 24 4mo ago A 38 tokens

Cenrax/researchswarm

Skill Claude CodeCodex

Verify that the generated code faithfully implements the algorithms and techniques described in the source research papers. Use this during code review to check paper-to-code alignment.

not rated 24 4mo ago A 30 tokens

pipeline-status

06

Cenrax/researchswarm

Skill Claude CodeCodex

Check the current state of the research-to-code pipeline by inspecting output directories for summaries, plans, code, and reviews. Use this when you need to understand what stages have been completed and what's pending.

not rated 24 4mo ago A 40 tokens

sandbox-execute

07

Cenrax/researchswarm

Skill Claude CodeCodex

Execute code in an isolated sandbox to verify correctness. Use this whenever you write or modify code to ensure it runs without errors before moving on.

not rated 24 4mo ago A 28 tokens

security-audit

08

Cenrax/researchswarm

Skill Claude CodeCodex

Perform a security audit on generated code checking for vulnerabilities, unsafe practices, and exposed secrets. Use this during code review to ensure the implementation is safe for production.

not rated 24 4mo ago A 32 tokens

summarize-paper

09

Cenrax/researchswarm

Skill Claude CodeCodex

Produce a structured, implementation-focused summary of a research paper. Use this when reading any paper (PDF, markdown, or arXiv) to ensure consistent, actionable summaries.

not rated 24 4mo ago A 35 tokens

write-tests

10

Cenrax/researchswarm

Skill Claude CodeCodex

Generate unit tests for implemented modules to verify correctness of individual components. Use this after writing code to create a test suite.

not rated 24 4mo ago A 24 tokens

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: