mega-minds

# Mega-Minds Critical Enforcement Rules **⚠️ MANDATORY COMPLIANCE - VIOLATION PENALTIES APPLY ⚠️** This document contains critical rules that MUST be followed for proper mega-minds operation. Violations of these rules will result in immediate corrective action and may cause system failures. --- ## Core System Rules ### Rule 1: Memory Safety Protocol 🧠 **ABSOLUTE MAXIMUM**: 2 concurrent agents (including orchestrator) **RATIONALE**: Prevents Claude Code memory crashes and ensures stable operation **ENFORCEMENT**: Automatic rejection of agent activation requests when limit exceeded **Required Actions**: - Check memory before activating agents: `npx mega-minds memory-status` - Save sessions between major phases: `npx mega-minds save-session "description"` - Compress context when approaching limits: `npx mega-minds compress-context` **VIOLATION PENALTY**: System will automatically deactivate agents to maintain stability --- ### Rule 2: Task Tool Enforcement 🔧 **MANDATORY**: ALL agent work MUST use Task tool **PROHIBITED**: Direct agent mentions like `@agent-name` without Task tool invocation **RATIONALE**: Ensures proper agent coordination and state tracking **Required Format**: ```markdown <function_calls> <invoke name="Task"> <parameter name="subagent_type">general-purpose</parameter> <parameter name="description">Brief task description</parameter> <parameter name="prompt">## Handoff to @[agent-name] [Complete handoff using communication-protocol.md template] </parameter> </invoke> </function_calls> ``` **Required Elements in Every Handoff**: 1. Visual activation marker: `🤖 @[agent-name] ACTIVE` 2. Complete handoff template from communication-protocol.md 3. Clear success criteria and acceptance conditions 4. Quality gate requirements before completion **VIOLATION PENALTY**: Direct mentions will be ignored; proper Task tool invocation required --- ### Rule 3: Project Orchestrator Constraints 🎯 **CRITICAL LIMITATION**: The @project-orchestrator-agent is LIMITED to coordination activities ONLY #### Orchestrator MUST DO: ✅ Plan and break down work into tasks ✅ Assign work to appropriate specialized agents using Task tool ✅ Coordinate handoffs between agents ✅ Track progress and manage todo lists ✅ Escalate conflicts to decision councils ✅ Ensure quality gates are followed #### Orchestrator MUST NOT DO: ❌ Write code, HTML, CSS, or configuration files ❌ Perform database operations or data analysis ❌ Create UI designs or user interfaces ❌ Execute tests or deployment procedures ❌ Make technical implementation decisions ❌ Bypass agent handoff protocols **VIOLATION PENALTY**: Any implementation work by the orchestrator MUST be immediately stopped and handed off to the appropriate specialist agent --- ### Rule 4: Quality Gate Compliance 🛡️ **MANDATORY**: Every piece of work must pass quality checks before progression **BLOCKING BEHAVIOR**: Failed gates prevent work from continuing **RATIONALE**: Ensures professional-grade output and prevents defects **Before Any Code Deployment**: - ✅ Code Review (by @code-review-agent) - ✅ Testing (by @testing-agent) - ✅ Security Scan (by @security-testing-agent) **Before Any Design Implementation**: - ✅ Architecture Approval (by @technical-architecture-agent) - ✅ User Experience Review (by @ux-ui-design-agent) **VIOLATION PENALTY**: Work progression blocked until all quality gates pass --- ### Rule 5: Agent Role Boundaries 🚧 **STRICT ENFORCEMENT**: Agents MUST NOT perform work outside their defined scope **RATIONALE**: Prevents role confusion and ensures expert-level output **REFERENCE**: See `workflows/agent-boundaries.md` for detailed boundaries **Example Violations**: - Frontend agent performing database operations - Testing agent writing production code - Design agent implementing security features **VIOLATION PENALTY**: Work must be transferred to appropriate specialist agent --- ## Communication Protocol Rules ### Rule 6: Mandatory Handoff Protocol 📋 **REQUIREMENT**: All agent transitions must follow communication-protocol.md templates **COMPONENTS**: - Proper context transfer - Clear task definition - Success criteria specification - Quality gate requirements **VIOLATION PENALTY**: Handoff rejection until proper protocol followed --- ### Rule 7: Sequential Phase Management ⏭️ **STRATEGY**: Sequential phases with mandatory checkpoints **REQUIRED FLOW**: Requirements → Design → Development → Testing → Deployment **CHECKPOINTS**: Session saves required between ALL phases **VIOLATION PENALTY**: Phase progression blocked until checkpoint completed --- ## Emergency Protocols ### Rule 8: Memory Crisis Response 🚨 **TRIGGER CONDITIONS**: - Memory usage >3.5GB - Multiple memory warnings - Claude Code performance degradation **IMMEDIATE ACTIONS**: 1. Save current session immediately 2. Compress context aggressively 3. Reduce concurrent agents to 1 4. Consider Claude Code restart if critical **AUTHORITY**: Any agent can trigger emergency protocol when memory critical --- ### Rule 9: Agent Failure Recovery 🔄 **FAILURE CONDITIONS**: - Agent non-responsiveness >5 minutes - Repeated handoff failures - Quality gate failures >3 attempts **RECOVERY PROCEDURE**: 1. Log failure details 2. Reset agent state 3. Re-initiate handoff with fresh context 4. Escalate to orchestrator if pattern continues --- ### Rule 10: System Override Authority 🔓 **OVERRIDE CONDITIONS**: - Critical production issues - Emergency bug fixes - System stability threats **AUTHORIZED PERSONNEL**: - Project orchestrator agent - System administrators - Emergency response teams **PROCEDURE**: 1. Document override reason 2. Implement minimum viable fix 3. Return to normal protocols ASAP 4. Post-incident review required --- ## Compliance Monitoring ### Violation Detection 🕵️ - Automated monitoring of agent behavior - Real-time rule compliance checking - Violation logging and reporting - Pattern analysis for repeat violations ### Enforcement Actions ⚖️ **Level 1 - Warning**: Rule reminder and guidance **Level 2 - Correction**: Forced compliance with proper procedure **Level 3 - Blocking**: Prevention of non-compliant actions **Level 4 - Reset**: Agent state reset and handoff restart ### Escalation Path 📈 1. Automated compliance system 2. Project orchestrator agent 3. System administrators 4. Emergency response protocols --- ## Rule Update Process 🔄 ### Version Control - All rule changes tracked with semantic versioning - Change log maintained with rationale - Backward compatibility considerations - Migration guides for major changes ### Approval Process - Technical review required for all changes - Impact assessment on existing workflows - Testing with representative scenarios - Staged rollout with monitoring ### Communication - Rule changes communicated to all agents - Training updates for complex changes - Documentation updates across all files - User notification for critical changes --- ## Quick Reference Card 🎯 ### Critical Commands ```bash npx mega-minds memory-status # Check before agent activation npx mega-minds save-session # Required between phases npx mega-minds compress-context # When memory warnings appear npx mega-minds agent-status # Monitor concurrent agents ``` ### Critical Limits - **Max Concurrent Agents**: 2 (including orchestrator) - **Memory Threshold**: <3.5GB before emergency protocols - **Handoff Timeout**: 5 minutes before failure escalation - **Quality Gate Retries**: 3 attempts before escalation ### Emergency Contacts - **System Issues**: Check `workflows/compliance-enforcement.md` - **Rule Violations**: See violation handling procedures - **Technical Support**: Contact system administrators - **Critical Failures**: Activate emergency response protocols --- *Last Updated: January 2025* *Version: 2.1.0* *Next Review: After major system updates* *Authority: Mega-Minds Development Team* **⚠️ REMEMBER: These rules exist to ensure system stability, quality output, and successful project completion. When in doubt, follow the rules strictly rather than taking shortcuts. ⚠️**