Install Agent Skills in Cursor
Cursor skills are agent packages that plug into the editor's agent loop, from pointing it at a repo and its constraints through to turning a working tree into a reviewable change. The sub-topics here map to that loop: setup, editing code toward a stated outcome, longer chat-and-agent runs that plan, change, and explain, terminal work where the agent runs the commands you would run locally, verification before you commit, and Git and PR handling. Use the ranked list by reading each skill's description against the step you are actually stuck on, then picking the one whose scope matches — a setup skill will not help you review a diff, and a PR skill assumes the change already exists. Start with the step closest to your current bottleneck rather than installing several at once.
8 to start with13 top options
Start with these.
Cursor is an editor with an agent: rules and context first, then edits, terminal, and review.
These are the closest matches for this work. Start with the first one and open it to install.
- Rank 01View Agent SkillPossible
thermo-nuclear-code-quality-review
Runs an extremely strict maintainability review focused on abstraction quality, giant files, and spaghetti-condition growth.
20,459 · popularity - Rank 02View Agent SkillPossible
deslop
Checks the diff against main and removes AI-generated code slop such as unnecessary comments, defensive checks, and any casts.
8,077 · popularity - Rank 03View Agent SkillPossible
agent-pulse
CLI that inspects local AI-agent activity across many tools including Cursor, reporting sessions, tokens, tool calls, and estimated cost.
47,517 · popularity - Rank 04View Agent SkillPossible
agent-browser
Fast browser automation CLI for AI agents using Chrome/Chromium via CDP with accessibility-tree snapshots and element refs.
985,975 · popularity - Rank 05View Agent SkillPossible
git-guardrails-claude-code
Sets up Claude Code PreToolUse hooks that block dangerous git commands like push, reset --hard, clean, and branch -D.
444,461 · popularity - Rank 06View Agent SkillPossible
safe-refactor
Restructures code while preserving behavior, bracketing structural edits with verification and keeping intermediate states buildable.
88,606 · popularity - Rank 07View Agent SkillPossible
plankton-code-quality
Sets up write-time code quality enforcement with formatters, linters, and auto-fix hooks on every file edit.
8,943 · popularity - Rank 08View Agent SkillPossible
gh-address-comments
OfficialHelp address review/issue comments on the open GitHub PR for the current branch using gh CLI; verify gh auth first and prompt the user to authenticate if not logged in.
Explore 5 more Agent Skills
- Rank 09View Agent SkillPossible
typescript-best-practices
TypeScript best practices covering discriminated unions, branded types, boundary validation, exhaustiveness, and type-system discipline.
5,876 · popularity - Rank 10View Agent SkillPossible
unslop
Edits text to remove AI-generated writing patterns, with a checklist of content, language, style, and jargon tells to fix.
21,157 · popularity - Rank 11View Agent SkillPossible
playwright
OfficialUse when the task requires automating a real browser from the terminal (navigation, form filling, snapshots, screenshots, data extraction, UI-flow debugging) via `playwright-cli` or the bundled wrapper script.
- Rank 12View Agent SkillPossible
setup-pre-commit
NotableSet up Husky pre-commit hooks with lint-staged (Prettier), type checking, and tests in the current repo. Use when user wants to add pre-commit hooks, set up Husky, configure lint-staged, or add commit-time formatting/typechecking/testing.
358,348 · popularity - Rank 13View Agent SkillPossible
verify-and-stop
NotableProve existing work meets acceptance conditions without expanding scope. Use for validation-only tasks, completion checks, focused gate runs, and last-mile proof.
54,721 · popularity
What this topic covers
Each sub-topic groups Agent Skills for a specific situation.
- Setup
- Point the agent at this repo, its tools, and its constraints.
- Edit code
- Ask it to change files toward a stated outcome.
- Chat & agent
- A longer loop than a single edit: plan, change, and explain.
- Terminal
- Let it run the same commands you would run locally.
- Verify
- Check the change the way you would before committing.
- Git & PRs
- Turn the working tree into a reviewable change.
Frequently asked questions
- How is this list ordered?
- It is ranked by observed search demand. Direct matches for the topic come first, and any single publisher is capped so one source cannot fill the page.
- Where should I start if I have not set Cursor up on this repo yet?
- Start with the setup step: rules and context first, so the agent knows the repo, its tools, and its constraints. Edits, terminal use, and review come after that foundation.
- Do I need separate skills for editing, terminal, and Git?
- They cover different parts of the loop. Editing changes files toward a stated outcome, terminal lets the agent run the commands you would run locally, and Git and PR skills turn the working tree into a reviewable change.
- What does a chat-and-agent skill do that a single edit does not?
- It runs a longer loop: plan, change, and explain. That suits work where the outcome needs reasoning across steps rather than one file change.
How to install an Agent Skill
Start with the first skill on this page. Claude Code reads a SKILL.md once it is in the skills directory:
- Claude Code — copy the SKILL.md for thermo-nuclear-code-quality-review into ~/.claude/skills/thermo-nuclear-code-quality-review/ for yourself, or .claude/skills/thermo-nuclear-code-quality-review/ to share it with a project.
- From the registry — run npx skills add cursor/plugins@thermo-nuclear-code-quality-review.
- Or download thermo-nuclear-code-quality-review from its source and place that SKILL.md in the directory above.
How to choose Check the source and the skill’s own instructions before you install. Official and Notable describe where a skill comes from.