AI-research-feedback skill
A skill set for academic work: paper review, pre-analysis plan check, grant proposal review, code audit and slide decks.
- 560 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Multi-agent runs spend quota and read the paper and code sources of the project.
- The typo-check skill offers to apply edits to `.tex` files: review the diff before accepting.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/ai-research-feedbackDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Multi-agent runs spend quota and read the paper and code sources of the project.
- The typo-check skill offers to apply edits to `.tex` files: review the diff before accepting.
This is third-party code. Review the repository files before installing.
What it does
AI-research-feedback is thirteen Claude Code skills for researchers. The review skills read a document and write a structured report: `review-paper` runs eight agents for a referee-style review, `review-paper-light` quickly checks contribution and identification, `review-paper-checks` hunts typos and inconsistencies, and there are checks for pre-analysis plans and grant proposals. Other skills audit analysis code (`review-paper-code`, `audit-analysis`), explain changes, prepare short paper versions and slide decks in Quarto, Beamer and PowerPoint. `pdf-to-markdown` converts PDFs to text. Reports are saved in the project `reviews/` folder with the date in the filename.
Who it is for. Economists and other researchers preparing a paper for submission who want independent text and code checks before reviewers see it.
Good fit when
- A draft is ready and a pre-submission review should list its weak spots.
- A pre-analysis plan or grant proposal is checked against registry or funder requirements.
- Analysis code is audited: does it produce the numbers the paper reports?
- A paper becomes a policy brief or slides with live figures.
- Spelling and internal references are proofread before submission with no editorial changes.
Not a fit when
- You need one short paragraph fix: a multi-agent run is too heavy.
- The work is not academic, for example marketing copy.
- Your plan limit is small: the full review launches eight agents.
Example request
Run /review-paper-light on draft.tex and show the main concerns about contribution and identification.Limitations
The review skills run only by typing `/name`, since Claude must not start a multi-agent check by itself. A full review spends a large share of quota. Reports go to `reviews/` and the skills deliberately ignore old drafts and earlier reports. The slide-deck skills and `pdf-to-markdown` may be picked up by the agent on its own. Beamer and PowerPoint decks need LaTeX and PowerPoint itself. The material is tuned to empirical economics.
How to disable. Delete the folders you no longer want from `~/.claude/skills/` (for example `review-paper`) and restart Claude Code. Reports in `reviews/` stay in the project.
FAQ
How do I install one skill?
Copy its folder from `Skills/` to `~/.claude/skills/<name>`. All but the three deck skills are a single `SKILL.md`.
Why do skills not run by themselves?
The review skills set `disable-model-invocation: true`, so a multi-agent run starts only on an explicit command.
Where do reports go?
To the project `reviews/` folder, for example `PRE_SUBMISSION_REVIEW_[date].md`. A rerun adds a `-v2` suffix.
How does light differ from the full review?
Light runs two agents and checks contribution, identification and causal overclaiming. The full review uses eight agents.
What about the decks?
Three skills build one deck in Quarto reveal.js, LaTeX Beamer and PowerPoint, with figures the presenter steps through.
Discussion
No questions or reviews yet.
Related
SkillsA skill turns an ML research repository into a conference paper draft in LaTeX with a claim-evidence map.
A marketplace of five Claude Code plugins for researchers: literature search, research design, project memory and manuscript work
Takes an agent from a topic to an arXiv review paper in LaTeX IEEEtran: approved plan, every citation checked against a live source, PDF build
A skill for master and PhD students that runs research through novelty and reality gates, from review to claim checks
Seven skills for planning research, reviewing literature, preparing experiment code and drafting a paper with traceable sources
Claude Code skills for the academic pipeline: research, paper drafting, peer review and citation checks
README in short
The README groups skills into text feedback, code feedback, communicating results and utilities, with install options and a command table listing agent counts and report names.