Pre-action gates that block AI coding agents from repeating known mistakes. Dual-memory recall (MemAlign-inspired principles + episodic context), auto-promoted prevention rules, and PreToolUse enforcement hooks.
mcp developer-tools cursor feedback-loop zero-config ai-agents guardrails llmops ai-memory model-context-protocol mcp-server agent-orchestration agent-memory cursor-plugin claude-code mcp-gateway agent-reliability pre-action-gates ai-agent-safety prevention-rules
-
Updated
Mar 24, 2026 - JavaScript