review-r — Claude Code skill: what it does, how to install
Runs a read-only review of R scripts to check tidyverse conventions, reproducibility, and code quality. Generates markdown reports without modifying source files.
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a claude-codeWhat review-r does
This skill delegates static analysis of .R scripts to the r-reviewer subagent. You can pass a specific file name, a lecture tag like LectureN, or the all keyword to scan both scripts/R/ and Figures/*/. The agent inspects code against rules stored in .claude/rules/r-code-conventions.md.
Tools like Edit and MultiEdit are disallowed. Each review is written directly to quality_reports/[script_name]_r_review.md, followed by a terminal summary highlighting the top three critical issues. Your source code stays untouched.
The skill does not execute code. For validating numerical outputs, run /audit-reproducibility instead.
What it helps with
- audit R scripts for tidyverse conventions before paper submission
- inspect research analysis files without altering source code
- generate structured markdown review reports in
quality_reports/ - surface critical bugs and reproducibility flaws across project folders
Example prompts for the agent
- review this R script
analysis.R - check the R code for
Lecture2 - audit the analysis code with argument
all
How to install review-r
1Skills CLI (Vercel)
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a claude-code2Manually: put the files in place
~/.claude/skills/review-r/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/review-r/
1Skills CLI (Vercel)
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a codex2Manually: put the files in place
.agents/skills/review-r/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a cursor2Manually: put the files in place
.cursor/skills/review-r/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install pedrohcgs/claude-code-my-workflow review-r2Manually: put the files in place
.github/skills/review-r/SKILL.md1Skills CLI (Vercel)
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a gemini-cli2Platform CLI
gemini skills install https://github.com/pedrohcgs/claude-code-my-workflow --path .claude/skills/review-rManually: put the files in place
.devin/skills/review-r/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about review-r
How to install review-r in Claude Code?
Run npx skills add pedrohcgs/claude-code-my-workflow --skill review-r -a claude-code.
Does review-r modify existing R files?
No, the skill operates in read-only mode. Editing tools like Edit and MultiEdit are disabled, and findings are written to markdown files in quality_reports/.
Does review-r run the code to verify results?
No, it does not execute scripts. If you need numerical verification of your analysis outputs, run /audit-reproducibility instead.
Source text
SKILL.md · MIT
--- name: review-r description: Read-only R code review protocol for `.R` scripts. Checks code quality, reproducibility, domain correctness, tidyverse idioms, and professional standards; produces a report without editing. Use when user says "review this R script", "check the R code", "audit the analysis code", "code review on the R", or when an R file is touched as part of a paper submission. NOT for running the code — pair with `/audit-reproducibility` for numeric verification. argument-hint: "[filename or 'all' or 'LectureN']" allowed-tools: ["Read", "Grep", "Glob", "Write", "Agent", "Task"] disallowed-tools: ["Edit", "MultiEdit"] --- # Review R Scripts Run the comprehensive R code review protocol. ## Steps 1. **Identify scripts to review:** - If `$ARGUMENTS` is a specific `.R` filename: review that file only - If `$ARGUMENTS` is `LectureN`: review all R scripts matching that lecture - If `$ARGUMENTS` is `all`: review all R scripts in `scripts/R/` and `Figures/*/` 2. **For each script, launch the `r-reviewer` agent** with instructions to: - Follow the full protocol in the agent instructions - Read `.claude/rules/r-code-conventions.md` for current standards - Return its report as its final response (ending with the findings `json` block); this skill saves it to `quality_reports/[script_name]_r_review.md` — the agent is read-only 3. **After all reviews complete**, present a summary: - Total issues found per script - Breakdown by severity (Critical / High / Medium / Low) - Top 3 most critical issues 4. **Leave the R source files unchanged** — the report goes to the user, who decides what to fix. Only produce reports. Fixes are applied after user review.
© pedrohcgs · MIT · pedrohcgs/claude-code-my-workflow
