Claude Code plugin that learns from your corrections — approved lessons become rules that earn autonomy through real use: verified → trusted → autonomous. One mistake = one level down. No daemon, no cloud, 100% local.
python skills memory ai-agents local-first self-improving anthropic agent-tools claude-code claude-code-plugin claude-plugin verified-ai trust-gradient compounding-trust no-daemon
-
Updated
Jun 22, 2026 - Python