Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
NeuroArxiv helps AI researchers and engineers validate technical designs by scanning arXiv for existing solutions before implementation begins. It fetches category-specific papers, analyzes them in parallel, and clusters findings to identify state-of-the-art protocols.
The process converges on a single recommended path supported by citations and documented pitfalls, preventing the redundant reconstruction of existing algorithms or systems techniques.
WHO IT'S FOR
ML / AI researchers
grounding new model architectures in prior art
Systems engineers
researching state-of-the-art protocols and techniques
Technical architects
validating design decisions against published research
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Follow the documented setup, then try a first task.
I’m using
Auto-detect installed agent
Installs: Bundled NeuroArxiv skill; the installer targets every detected supported agent and defaults to Claude Code if neither agent configuration directory is found. · Claude Code, Codex CLI
Before you start
Node.js 18 or newer
Run this terminal command. It requires no repository clone or separate build step and automatically selects supported agents present on the machine.
npx github:UditAkhourii/neuroarxiv install
Restart the installed agent or begin a new session. In Claude Code, invoke NeuroArxiv with `/neuroarxiv "<problem>"`, replacing `<problem>` with the technical problem to research. In Codex CLI, ask for the neuroarxiv skill by name.
Start a new Codex session, then ask for the neuroarxiv skill by name. If CODEX_HOME is configured, the installer honors it instead of the default Codex configuration directory.
A description of the technical mechanism or architecture decision
Relevant scale, workload, system, and design constraints
Why the decision will require significant effort or be expensive to reverse
Use the neuroarxiv skill to check arXiv prior art for this open-ended technical mechanism: [describe the architecture, algorithm, ML/systems technique, or protocol]. Context and constraints: [describe scale, workload, requirements, and existing system]. Explain what prior work suggests, then recommend one path with citations, a practical first step, and known prior-art pitfalls. Do not implement anything yet.
Claude
Codex