Skip to content

smith-xyz/cursor-plugins

Repository files navigation

Cursor-Plugins

Personally curated commands, skills, and subagents for Cursor or Claude Code.

Mental model

Layer Use for
Skills (skills/<name>/) Procedures, style, and stack conventions. Primary place for repeatable patterns and scripts.
Commands (commands/*.md) Slash playbooks with a clear outcome (scripts, gh, scaffold). Avoid duplicating full skill bodies here.
Agents (agents/*.md) Separate context, optional readonly. Point at skills so delegated work follows your conventions.

Flow: Skills hold the substance; agents load those skills for bounded tasks; commands stay thin and outcome-focused.

Layout

  • skills/<name>/ — Instructions, scripts, assets.
  • commands/ — Slash playbooks for named outcomes only.
  • agents/Subagents (separate context, optional readonly). Installed to ~/.cursor/agents/ (and ~/.claude/agents/ when using the claude vendor). Invoke with /name.

Slash commands (outcomes)

Command Purpose
scaffold New project from templates (project-scaffold skill)
work-review Summarize work from transcripts; optional Jira (work-review skill)
find-issues Find GitHub issues for OSS contribution (gh)
reproduce-issue Reproduce a GitHub issue / build a test case (gh)

Stack and domain work (Go, React, Python, TypeScript, Rust, OpenShift) lives in skills and subagents (/go-dev, /react-dev, etc.), not duplicate slash commands.

Deps

Homebrew required.

make install-deps

Installation

Copies rules/, commands/, skills/, and agents/ into ~/.$vendor/.

make install

After make install, run make lint on changed files if you edit markdown.

About

My cursor plugins

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors