AI configs that actually work.
Every rule, MCP server, and skill is quality-scored, security-scanned, and verified across environments. Install in one command.
Works with your tools
How it works
Browse and search
Find rules, MCPs, skills, and more for your environment.
Check quality scores
Every asset is measured on token efficiency, security, and reliability.
Install with one command
Copy the install command. Works with Cursor, Windsurf, Claude, and more.
Quality is measured, not voted.
Five automated signals replace star ratings and download counts.
Token efficiency
Measures actual token consumption, not self-reported estimates.
Install success rate
Tracked across real CLI installations.
Schema cleanliness
JSON and YAML validation against published standards.
Freshness
Penalizes stale assets. Rewards active maintenance.
Security scan
VirusTotal, Semgrep, and sandboxed execution.
Example: Senior Cursor Rules
Built on verified trust
VirusTotal scanning
Every upload scanned against 70+ antivirus engines.
Semgrep analysis
Static analysis catches security anti-patterns before publishing.
Verified-install reviews
Only users who installed an asset via CLI can leave reviews.
What people are building
Semantic search over local files
Embedding-based file search rather than just grep.
CursorRules for Rust
I maintain a private Rust .cursorrules. Would anyone use it?
Best practices for composing multiple rulesets?
I have three .cursorrules files. What is the recommended way to layer them?
GitHub Actions MCP integration
Trigger CI from Claude Code via the MCP server.
MCP server security best practices
Security considerations for exposing MCP servers?
Stop guessing. Start building.
50+ assets scored on real signals. Install what works, skip what doesn't.