Added Nov 3, 2026
Open Research Skills
Ten everyday research Skills: figures, papers, grants, novelty checks, mock reviews and getting code ready to release.
- Ecosystem
- Agent Skills
- Skill category
- Research
- Creator / organization
- gxCaesar、shapsider、hehh77(MAINTAINERS 所列;仓库属于 GitHub 用户 gxCaesar)
What it is
Open Research Skills is a research toolbox maintained by gxCaesar, shapsider and hehh77; the README calls it an agent skill toolbox for everyday research. It has 10 Skills you can install one at a time: research figures, conference papers, journal papers, grant proposals, an end-to-end research pipeline, method iteration, novelty checks, cold review, invited peer review, and prep work before releasing code and data. The research pipeline Skill says it has no remote launcher and no automatic submission, and that publishing, submitting or spending on training needs your separate approval. The site lists the whole repo as one entry.
Who it's for:Researchers writing papers and grants, preparing submissions, or getting code ready to share.
Before you start: In Claude Code, two plugin commands install the whole set. For Codex and Antigravity, copy the full Skill folders or run bash scripts/install_skills.sh locally. That script skips folders that already exist by default; with --force it deletes the old folder and copies over it, so back up first. You need Python 3.9 or later, and features like figures have their own dependencies. It reads your papers, grant drafts, data and code, and writes papers, figures and release packages on your machine; figures use the image model in your current session. The README and tutorials are in Chinese. Apache-2.0 license.
Based on the project's public documentation. Not tested by AI Inspiration World.