Kevin-Liu-01/loop
self-improving skills. stay in the loop
self-improving skills. stay in the loop
npx skills add Kevin-Liu-01/loopself-improving skills. stay in the loop
Sentinel — a read-only, pre-action on-chain risk gate. Given (address, action) it returns a safe/caution/dangerous verdict before an agent moves value. Packaged as a Pharos Skill Engine Step-0 pre-check (MCP server + SKILL.md/CLI).
Provider-neutral Agent Skill for designing, auditing, and operating multi-agent teams
Codex Skill and CLI orchestration layer for safer agentic coding.
Agentic CLI tool skills for Claude Code — 7 domain-grouped skills covering 26 CLI tools
Virtual environments for AI coding harness configs — named, composable, version-controlled bundles of CLAUDE.md, .cursorrules, skills, MCP entries, and other agent configuration. Think Python's venv, but for the rules that shape how AI coding agents behave.
Naoray's personal Claude Code, Cursor & Codex skills. Connect in one command with scribe.