Cursor Rules Pack
A pack of 10 Cursor project rules (.mdc files in .cursor/rules) that stop common AI coding mistakes: guessed imports, any types, "use client" on whole pages, unauthenticated API routes and weakened tests. Each rule is scoped with globs or a description so it only loads when relevant.
Who it's for
Cursor users shipping TypeScript, Next.js or Python apps who want the Agent to follow the same standards every time.
What goes in
- Cursor
- A project repository
What comes out
- An always-on core workflow rule
- File-scoped rules for TypeScript, Next.js App Router, React, Python and Tailwind
- On-demand rules for testing, API security, git commits and Supabase
Setup in 4 steps
- Unzip and run tools/install.sh /path/to/repo (or pick individual rules).
- Open Cursor → Customize → Rules to confirm they're listed.
- Ask the Agent for a change and watch it follow the rules.
- Delete or edit rules to match your stack, then commit .cursor/rules.
Tool stack
Cost per run
No extra cost. Applied rules add roughly 100 to 250 tokens each to the Agent's context.
What's included
- 10 Cursor project rules (.mdc)
- Installer with backups + validator + AGENTS.md
- README setup guide
- Offline sample data + tests
- Honest cost notes
- Changelog
FAQ
What are Cursor rules?
Project rules are .mdc files in .cursor/rules with frontmatter (description, globs, alwaysApply) that Cursor adds to the Agent's context, always, for matching files, or when the Agent decides they're relevant.
Do these rules affect Cursor Tab autocomplete?
No. Cursor applies rules to Agent (Chat), not to Tab completions.
Can I use only some of the rules?
Yes. The installer takes rule names, for example install.sh ./repo core-workflow typescript.
Is this the same as AGENTS.md?
AGENTS.md is a single plain-markdown alternative. These .mdc rules can be scoped per file type, which keeps context smaller.