CouchLoop EQ

Universal entry point for CouchLoop. Routes ANY command to the right tool. ALWAYS use for ambiguous/loose commands. Handles: - Sessions: "end session", "start", "done", "wrap up", "goodbye", "resume", "where should I start", "hi", "hey" - Status: "how am I doing", "what do you know about me", "show my progress", "my settings" - Memory: "save", "remember", "checkpoint", "recall", "don't forget" - Code: "review code", "check this", "find bugs", "is this safe" - Packages: "audit dependencies", "outdated", "npm audit", "upgrade" - Protection: "backup", "freeze code", "rollback", "undo", "restore" - Journeys: "I'm stressed", "feeling anxious", "help me", "need to talk" - Verification: "verify this", "check my response", "is this correct", "does this package exist" - Brainstorm: "brainstorm", "think through", "map out feature", "help me design", "I have an idea", "flesh out", "trade-offs" This tool should be invoked for ANY ambiguous or loose command related to sessions, memory, code quality, verification, brainstorming, or guided self-reflection.

CRITICAL: Pre-delivery verification for AI-generated content. Call BEFORE presenting code, package recommendations, or factual claims to users. Catches: hallucinated packages (24% of AI suggestions don't exist!), hallucinated APIs, incorrect imports, ESM/CJS confusion, deprecated methods, security vulnerabilities, inconsistencies with previous statements. Returns verification result with fixes if needed.

Dashboard and quick status checks. Use for: "how am I doing", "what's my progress", "show my history", "what do you know about me", "my settings", "context window", "backup status". Returns personalized status with actionable next steps.

Start or continue an AI conversation with built-in crisis detection, guided self-reflection journeys, and session memory. Includes brainstorm mode for dev ideation. Triggers: "end session", "start session", "wrap up", "done for now", "talk", "chat", "feeling", "stressed", "help me", "brainstorm", "think through", "map out feature".

Dev thinking partner for architecture decisions, feature design, trade-offs, and technical exploration. Asks reflective questions to help you arrive at your own best solution, then provides concrete analysis when you've narrowed options. Triggers: "brainstorm", "think through", "map out", "help me design", "I have an idea", "flesh out", "trade-offs", "pros and cons", "should I use X or Y".

Complete code review: security vulnerabilities (SQL injection, XSS, secrets), code quality (console.logs, TODOs, error handling), code smells (complexity, bloat), and AI-generated errors (hallucinated APIs, build context issues). One call, full analysis. Triggers: "review", "check code", "analyze", "security check", "lint", "find bugs", "is this safe".

Complete dependency audit: validates packages exist and are legitimate (catches typosquatting), checks for outdated versions and security vulnerabilities, generates upgrade reports with migration guides and breaking changes. Triggers: "audit", "check dependencies", "outdated", "vulnerable packages", "upgrade", "npm audit", "security scan".

Capture and preserve important context from conversations. Automatically routes to the right storage: checkpoints for progress, insights for realizations, context for technical decisions. Prevents AI amnesia across conversations. Triggers: "save", "remember this", "checkpoint", "note", "don't forget", "keep track", "save progress", "log this".

File protection and safety: prevent accidental deletions, create automatic backups, rollback changes, enable code freeze mode. Essential for safe AI-assisted development. Triggers: "backup", "protect", "freeze", "rollback", "undo", "restore", "safe mode".