{
  "package": {
    "name": "agentic-flow",
    "version": "2.0.3",
    "description": "Production-ready AI agent orchestration platform with 66 specialized agents, 213 MCP tools, ReasoningBank learning memory, and autonomous multi-agent swarms. Built by @ruvnet with Claude Agent SDK, neural networks, memory persistence, GitHub integration.",
    "repository": "https://github.com/ruvnet/agentic-flow",
    "author": "ruv (https://github.com/ruvnet)",
    "license": "MIT"
  },
  "capabilities": [
    {
      "name": "Multi-Agent Swarm Orchestration",
      "description": "Orchestrate multi-agent swarms with mesh, hierarchical, ring, star, and adaptive topologies for parallel task execution and intelligent coordination",
      "keywords": ["swarm", "multi-agent", "orchestration", "coordination", "topology", "mesh", "hierarchical", "parallel"],
      "category": "swarm",
      "example_prompts": ["Initialize a swarm with hierarchical topology", "Spawn 5 agents to work in parallel", "Coordinate multiple agents on a complex task", "Set up agent swarm for code review"]
    },
    {
      "name": "AgentDB Vector Search",
      "description": "High-performance vector database with HNSW indexing (150x-12,500x faster), quantization (4-32x memory reduction), and sub-millisecond search",
      "keywords": ["vector", "search", "HNSW", "embeddings", "similarity", "semantic", "quantization", "database"],
      "category": "memory",
      "example_prompts": ["Search for similar documents in the knowledge base", "Find code patterns matching this query", "Initialize vector database with binary quantization", "Query vectors with cosine similarity"]
    },
    {
      "name": "ReasoningBank Learning Memory",
      "description": "Adaptive learning system for pattern recognition, strategy optimization, and continuous improvement with persistent memory",
      "keywords": ["learning", "memory", "patterns", "reasoning", "adaptive", "experience", "strategy"],
      "category": "learning",
      "example_prompts": ["Learn from this successful approach", "Find optimal strategy for this task", "Store this pattern for future use", "Retrieve similar past experiences"]
    },
    {
      "name": "Reinforcement Learning Plugins",
      "description": "9 RL algorithms: Decision Transformer, Q-Learning, SARSA, Actor-Critic, Active Learning, Adversarial Training, Curriculum Learning, Federated Learning, Multi-Task Learning",
      "keywords": ["reinforcement-learning", "RL", "decision-transformer", "q-learning", "sarsa", "actor-critic", "training"],
      "category": "learning",
      "example_prompts": ["Create a decision transformer plugin", "Train agent using Q-learning", "Set up actor-critic for continuous control", "Enable curriculum learning for complex tasks"]
    },
    {
      "name": "Flash Attention",
      "description": "Optimized attention mechanism with 2.49x-7.47x speedup and 50-75% memory reduction",
      "keywords": ["attention", "flash-attention", "performance", "optimization", "speedup", "memory"],
      "category": "performance",
      "example_prompts": ["Enable flash attention for faster inference", "Optimize attention with memory reduction", "Configure 8-head attention mechanism"]
    },
    {
      "name": "SONA (Self-Optimizing Neural Architecture)",
      "description": "Neural architecture with <0.05ms adaptation overhead, automatic optimization, and continuous improvement",
      "keywords": ["SONA", "neural", "self-optimizing", "adaptation", "architecture", "learning"],
      "category": "neural",
      "example_prompts": ["Enable SONA for self-optimizing agent", "Configure neural adaptation rate", "Train SONA model on task patterns"]
    },
    {
      "name": "MCP Server Integration",
      "description": "213 MCP tools for Claude Code integration including agent management, memory operations, neural training, and GitHub integration",
      "keywords": ["MCP", "tools", "Claude", "integration", "server", "fastmcp"],
      "category": "integration",
      "example_prompts": ["Start MCP server for Claude Code", "Add agentic-flow to Claude Code", "Use MCP tools for agent coordination"]
    },
    {
      "name": "Hive-Mind Consensus",
      "description": "Byzantine fault-tolerant consensus with queen-led coordination, supporting raft, gossip, CRDT, and quorum protocols",
      "keywords": ["consensus", "hive-mind", "byzantine", "raft", "gossip", "CRDT", "distributed"],
      "category": "coordination",
      "example_prompts": ["Initialize hive-mind consensus", "Set up Byzantine fault-tolerant coordination", "Enable raft consensus for leader election"]
    },
    {
      "name": "QUIC Synchronization",
      "description": "Sub-millisecond latency synchronization between AgentDB instances with automatic retry, multiplexing, and TLS 1.3 encryption",
      "keywords": ["QUIC", "sync", "distributed", "latency", "transport", "encryption"],
      "category": "distributed",
      "example_prompts": ["Enable QUIC sync between database nodes", "Configure distributed AgentDB cluster", "Set up cross-node synchronization"]
    },
    {
      "name": "Agent Booster",
      "description": "352x faster code editing with AST-based transformations for simple operations (var-to-const, add-types, remove-console)",
      "keywords": ["agent-booster", "AST", "transform", "code-editing", "fast", "optimization"],
      "category": "performance",
      "example_prompts": ["Use agent booster for simple code transform", "Convert var to const across files", "Add TypeScript types automatically"]
    },
    {
      "name": "Background Workers (12 Types)",
      "description": "Background workers for ultralearn, optimize, consolidate, predict, audit, map, preload, deepdive, document, refactor, benchmark, and testgaps",
      "keywords": ["workers", "background", "async", "optimization", "audit", "benchmark", "documentation"],
      "category": "automation",
      "example_prompts": ["Dispatch audit worker for security scan", "Run benchmark worker for performance", "Trigger testgaps worker for coverage analysis"]
    },
    {
      "name": "Hooks System (27 Hooks)",
      "description": "Lifecycle hooks for pre/post edit, command, task, session management, routing, intelligence, and worker dispatch",
      "keywords": ["hooks", "lifecycle", "events", "routing", "session", "automation"],
      "category": "automation",
      "example_prompts": ["Set up pre-task hook for coordination", "Enable post-edit hook for learning", "Configure session hooks for persistence"]
    },
    {
      "name": "GitHub Integration",
      "description": "PR management, code review swarms, issue tracking, release management, and workflow automation",
      "keywords": ["GitHub", "PR", "code-review", "issues", "release", "workflow", "automation"],
      "category": "integration",
      "example_prompts": ["Create PR with AI-generated description", "Run code review swarm on changes", "Manage GitHub issues with agents"]
    },
    {
      "name": "SPARC Methodology",
      "description": "Specification, Pseudocode, Architecture, Refinement, Completion methodology with specialized agents",
      "keywords": ["SPARC", "methodology", "specification", "architecture", "development"],
      "category": "methodology",
      "example_prompts": ["Start SPARC workflow for new feature", "Use SPARC specification agent", "Run architecture phase with SPARC"]
    },
    {
      "name": "Hyperbolic Embeddings",
      "description": "Poincare ball model embeddings for hierarchical data representation with custom distance metrics",
      "keywords": ["hyperbolic", "poincare", "embeddings", "hierarchical", "distance", "geometry"],
      "category": "embeddings",
      "example_prompts": ["Use hyperbolic embeddings for hierarchy", "Configure Poincare ball model", "Calculate hyperbolic distance"]
    },
    {
      "name": "EWC++ Continual Learning",
      "description": "Elastic Weight Consolidation to prevent catastrophic forgetting during continuous learning",
      "keywords": ["EWC", "continual-learning", "catastrophic-forgetting", "consolidation"],
      "category": "learning",
      "example_prompts": ["Enable EWC++ for continual learning", "Prevent forgetting with consolidation", "Configure elastic weight constraints"]
    },
    {
      "name": "LoRA Fine-Tuning",
      "description": "Low-Rank Adaptation for efficient model fine-tuning with 99% parameter reduction",
      "keywords": ["LoRA", "fine-tuning", "adaptation", "parameters", "efficient"],
      "category": "training",
      "example_prompts": ["Fine-tune model with LoRA", "Apply LoRA adaptation to agent", "Configure low-rank parameters"]
    },
    {
      "name": "GNN Query Refinement",
      "description": "Graph Neural Network based query refinement with +12.4% recall improvement",
      "keywords": ["GNN", "graph", "query", "refinement", "recall", "neural-network"],
      "category": "search",
      "example_prompts": ["Enable GNN query refinement", "Improve search with graph analysis", "Configure graph-aware retrieval"]
    }
  ],
  "cli_commands": [
    {"name": "init", "description": "Project initialization with wizard, presets, skills, and hooks configuration", "subcommands": ["--wizard", "--preset", "--skills", "--hooks"], "keywords": ["init", "setup", "project", "wizard"], "category": "core", "example_prompts": ["Initialize new agentic-flow project", "Run project setup wizard"]},
    {"name": "agent", "description": "Agent lifecycle management including spawn, list, status, stop, metrics, pool, health, and logs", "subcommands": ["spawn", "list", "status", "stop", "metrics", "pool", "health", "logs"], "keywords": ["agent", "spawn", "status", "lifecycle", "pool", "health"], "category": "agent", "example_prompts": ["Spawn a coder agent", "List all active agents", "Check agent health"]},
    {"name": "swarm", "description": "Multi-agent swarm coordination with init, status, shutdown, scale, and topology management", "subcommands": ["init", "status", "shutdown", "scale", "topology"], "keywords": ["swarm", "multi-agent", "coordination", "topology"], "category": "swarm", "example_prompts": ["Initialize swarm with mesh topology", "Check swarm status", "Scale swarm to 10 agents"]},
    {"name": "memory", "description": "AgentDB memory operations with vector search (150x-12,500x faster): store, search, list, retrieve, init, stats, export, import", "subcommands": ["store", "search", "list", "retrieve", "init", "stats", "export", "import", "delete", "vacuum", "merge"], "keywords": ["memory", "store", "search", "vector", "database", "AgentDB"], "category": "memory", "example_prompts": ["Store pattern in memory", "Search for similar patterns", "Export memory database"]},
    {"name": "mcp", "description": "MCP server management with start, stop, status, list tools, and tool execution", "subcommands": ["start", "stop", "status", "list", "call", "tools", "register", "unregister", "restart"], "keywords": ["MCP", "server", "tools", "integration"], "category": "integration", "example_prompts": ["Start MCP server", "List available MCP tools", "Call MCP tool"]},
    {"name": "task", "description": "Task creation, assignment, status tracking, and lifecycle management", "subcommands": ["create", "assign", "status", "complete", "cancel", "list"], "keywords": ["task", "create", "assign", "workflow"], "category": "task", "example_prompts": ["Create new task", "Assign task to agent", "Check task status"]},
    {"name": "session", "description": "Session state management with save, restore, list, delete, and info operations", "subcommands": ["save", "restore", "list", "delete", "info", "export", "import"], "keywords": ["session", "state", "persistence", "restore"], "category": "session", "example_prompts": ["Save current session", "Restore previous session", "List saved sessions"]},
    {"name": "config", "description": "Configuration management with get, set, list, reset, export, and import", "subcommands": ["get", "set", "list", "reset", "export", "import", "validate"], "keywords": ["config", "settings", "configuration"], "category": "config", "example_prompts": ["Get configuration value", "Set configuration option", "Export configuration"]},
    {"name": "hooks", "description": "Self-learning hooks system with 27 hooks and 12 background workers", "subcommands": ["pre-edit", "post-edit", "pre-command", "post-command", "pre-task", "post-task", "session-start", "session-end", "session-restore", "route", "explain", "pretrain", "build-agents", "metrics", "transfer", "list", "intelligence", "worker", "progress", "statusline", "coverage-route", "coverage-suggest", "coverage-gaps"], "keywords": ["hooks", "lifecycle", "learning", "workers", "automation"], "category": "hooks", "example_prompts": ["Run pre-task hook", "Dispatch background worker", "Check hook metrics"]},
    {"name": "hive-mind", "description": "Queen-led Byzantine fault-tolerant consensus with init, status, join, leave, consensus, and broadcast", "subcommands": ["init", "status", "join", "leave", "consensus", "broadcast"], "keywords": ["hive-mind", "consensus", "byzantine", "coordination"], "category": "consensus", "example_prompts": ["Initialize hive-mind", "Join agent to hive", "Broadcast message to hive"]},
    {"name": "daemon", "description": "Background worker daemon management with start, stop, status, trigger, and enable", "subcommands": ["start", "stop", "status", "trigger", "enable"], "keywords": ["daemon", "background", "worker", "service"], "category": "daemon", "example_prompts": ["Start background daemon", "Check daemon status", "Enable daemon worker"]},
    {"name": "neural", "description": "Neural pattern training with train, status, patterns, predict, and optimize", "subcommands": ["train", "status", "patterns", "predict", "optimize"], "keywords": ["neural", "training", "patterns", "predict", "optimize"], "category": "neural", "example_prompts": ["Train neural model", "View learned patterns", "Predict optimal approach"]},
    {"name": "security", "description": "Security scanning with scan, audit, cve, threats, validate, and report", "subcommands": ["scan", "audit", "cve", "threats", "validate", "report"], "keywords": ["security", "scan", "audit", "CVE", "threats"], "category": "security", "example_prompts": ["Run security scan", "Check for CVE vulnerabilities", "Generate security report"]},
    {"name": "performance", "description": "Performance profiling with benchmark, profile, metrics, optimize, and report", "subcommands": ["benchmark", "profile", "metrics", "optimize", "report"], "keywords": ["performance", "benchmark", "profile", "metrics", "optimize"], "category": "performance", "example_prompts": ["Run performance benchmark", "Profile component", "Generate performance report"]},
    {"name": "embeddings", "description": "Vector embeddings operations with embed, batch, search, and init (75x faster with ONNX)", "subcommands": ["embed", "batch", "search", "init"], "keywords": ["embeddings", "vector", "ONNX", "batch"], "category": "embeddings", "example_prompts": ["Generate embeddings for text", "Batch embed documents", "Search with embeddings"]},
    {"name": "doctor", "description": "System diagnostics with health checks for Node.js, npm, Git, config, daemon, memory, and API keys", "subcommands": ["--fix"], "keywords": ["doctor", "diagnostics", "health", "fix"], "category": "system", "example_prompts": ["Run system diagnostics", "Fix detected issues", "Check system health"]},
    {"name": "migrate", "description": "V2 to V3 migration with status, run, rollback, validate, and plan", "subcommands": ["status", "run", "rollback", "validate", "plan"], "keywords": ["migrate", "upgrade", "V3", "rollback"], "category": "migration", "example_prompts": ["Check migration status", "Run V3 migration", "Rollback migration"]}
  ],
  "agent_types": [
    {"name": "coder", "description": "Code implementation agent with pattern learning and best practices", "keywords": ["code", "implementation", "development", "programming"], "category": "development", "example_prompts": ["Write a REST API endpoint", "Implement the feature", "Fix this bug"]},
    {"name": "reviewer", "description": "Code review agent with pattern-based issue detection", "keywords": ["review", "code-quality", "analysis", "feedback"], "category": "development", "example_prompts": ["Review this pull request", "Check code quality", "Find potential issues"]},
    {"name": "tester", "description": "Test generation agent that learns from failures", "keywords": ["test", "testing", "QA", "coverage"], "category": "development", "example_prompts": ["Write unit tests", "Generate test cases", "Check test coverage"]},
    {"name": "planner", "description": "Task orchestration agent with MoE routing", "keywords": ["planning", "orchestration", "task", "coordination"], "category": "coordination", "example_prompts": ["Plan the implementation", "Break down this task", "Create project roadmap"]},
    {"name": "researcher", "description": "Enhanced pattern recognition agent for analysis", "keywords": ["research", "analysis", "patterns", "investigation"], "category": "research", "example_prompts": ["Research this topic", "Analyze codebase patterns", "Find best practices"]},
    {"name": "security-architect", "description": "Security architecture and threat modeling agent", "keywords": ["security", "architecture", "threats", "vulnerabilities"], "category": "security", "example_prompts": ["Design secure architecture", "Model potential threats", "Review security"]},
    {"name": "security-auditor", "description": "Security audit and CVE scanning agent", "keywords": ["audit", "CVE", "security-scan", "vulnerabilities"], "category": "security", "example_prompts": ["Audit security", "Scan for CVEs", "Check for vulnerabilities"]},
    {"name": "memory-specialist", "description": "Memory management and optimization agent", "keywords": ["memory", "optimization", "storage", "patterns"], "category": "optimization", "example_prompts": ["Optimize memory usage", "Manage agent memory", "Consolidate patterns"]},
    {"name": "performance-engineer", "description": "Performance optimization and profiling agent", "keywords": ["performance", "profiling", "optimization", "benchmarks"], "category": "optimization", "example_prompts": ["Optimize performance", "Profile application", "Find bottlenecks"]},
    {"name": "hierarchical-coordinator", "description": "Queen-worker coordination model agent", "keywords": ["coordinator", "hierarchical", "queen", "workers"], "category": "coordination", "example_prompts": ["Coordinate worker agents", "Manage task distribution", "Lead swarm"]},
    {"name": "mesh-coordinator", "description": "Peer consensus coordination agent", "keywords": ["mesh", "peer", "consensus", "distributed"], "category": "coordination", "example_prompts": ["Coordinate peer agents", "Reach consensus", "Distributed coordination"]},
    {"name": "adaptive-coordinator", "description": "Dynamic coordination mechanism selection agent", "keywords": ["adaptive", "dynamic", "coordination", "flexible"], "category": "coordination", "example_prompts": ["Adapt coordination strategy", "Dynamic task routing", "Flexible orchestration"]},
    {"name": "byzantine-coordinator", "description": "Byzantine fault-tolerant coordination agent", "keywords": ["byzantine", "fault-tolerant", "consensus", "reliable"], "category": "consensus", "example_prompts": ["Handle faulty agents", "Byzantine consensus", "Fault-tolerant coordination"]},
    {"name": "raft-manager", "description": "Raft consensus protocol manager agent", "keywords": ["raft", "consensus", "leader-election", "log-replication"], "category": "consensus", "example_prompts": ["Manage raft consensus", "Leader election", "Log replication"]},
    {"name": "gossip-coordinator", "description": "Gossip protocol coordination agent", "keywords": ["gossip", "epidemic", "eventual-consistency", "distributed"], "category": "consensus", "example_prompts": ["Spread information via gossip", "Eventual consistency", "Epidemic broadcast"]},
    {"name": "crdt-synchronizer", "description": "CRDT-based conflict-free synchronization agent", "keywords": ["CRDT", "conflict-free", "synchronization", "distributed"], "category": "consensus", "example_prompts": ["Sync with CRDTs", "Conflict-free updates", "Distributed state"]},
    {"name": "pr-manager", "description": "Pull request management agent", "keywords": ["PR", "pull-request", "GitHub", "review"], "category": "github", "example_prompts": ["Create pull request", "Manage PR lifecycle", "Review PR changes"]},
    {"name": "code-review-swarm", "description": "Multi-agent code review swarm", "keywords": ["code-review", "swarm", "review", "quality"], "category": "github", "example_prompts": ["Review code with swarm", "Multi-agent review", "Parallel code analysis"]},
    {"name": "issue-tracker", "description": "GitHub issue tracking agent", "keywords": ["issues", "tracking", "GitHub", "bugs"], "category": "github", "example_prompts": ["Track GitHub issues", "Create issue", "Manage issue lifecycle"]},
    {"name": "release-manager", "description": "Release management and versioning agent", "keywords": ["release", "versioning", "deployment", "changelog"], "category": "github", "example_prompts": ["Create release", "Generate changelog", "Manage versions"]},
    {"name": "workflow-automation", "description": "GitHub workflow automation agent", "keywords": ["workflow", "automation", "CI/CD", "GitHub-Actions"], "category": "github", "example_prompts": ["Automate workflow", "Create CI/CD pipeline", "Manage GitHub Actions"]},
    {"name": "sparc-coord", "description": "SPARC methodology coordinator agent", "keywords": ["SPARC", "methodology", "coordinator", "workflow"], "category": "methodology", "example_prompts": ["Coordinate SPARC workflow", "Run specification phase", "SPARC orchestration"]},
    {"name": "specification", "description": "SPARC specification writer agent", "keywords": ["specification", "requirements", "SPARC", "design"], "category": "methodology", "example_prompts": ["Write specification", "Define requirements", "Document constraints"]},
    {"name": "pseudocode", "description": "SPARC pseudocode generator agent", "keywords": ["pseudocode", "algorithm", "SPARC", "design"], "category": "methodology", "example_prompts": ["Generate pseudocode", "Design algorithm", "Write pseudocode spec"]},
    {"name": "architecture", "description": "SPARC architecture designer agent", "keywords": ["architecture", "design", "SPARC", "structure"], "category": "methodology", "example_prompts": ["Design architecture", "Create system design", "Architecture planning"]},
    {"name": "refinement", "description": "SPARC refinement and optimization agent", "keywords": ["refinement", "optimization", "SPARC", "improvement"], "category": "methodology", "example_prompts": ["Refine implementation", "Optimize solution", "Improve architecture"]},
    {"name": "backend-dev", "description": "Backend development specialist agent", "keywords": ["backend", "server", "API", "development"], "category": "development", "example_prompts": ["Build backend API", "Server development", "Database integration"]},
    {"name": "mobile-dev", "description": "Mobile development specialist agent", "keywords": ["mobile", "iOS", "Android", "React-Native"], "category": "development", "example_prompts": ["Build mobile app", "iOS development", "Android feature"]},
    {"name": "ml-developer", "description": "Machine learning development agent", "keywords": ["ML", "machine-learning", "AI", "models"], "category": "development", "example_prompts": ["Build ML model", "Train classifier", "ML pipeline"]},
    {"name": "cicd-engineer", "description": "CI/CD pipeline engineering agent", "keywords": ["CI/CD", "pipeline", "automation", "DevOps"], "category": "devops", "example_prompts": ["Setup CI/CD", "Build pipeline", "Automate deployment"]},
    {"name": "api-docs", "description": "API documentation writer agent", "keywords": ["API", "documentation", "OpenAPI", "Swagger"], "category": "documentation", "example_prompts": ["Document API", "Generate OpenAPI spec", "Write API docs"]},
    {"name": "system-architect", "description": "System architecture design agent", "keywords": ["system", "architecture", "design", "infrastructure"], "category": "architecture", "example_prompts": ["Design system architecture", "Infrastructure planning", "System design"]},
    {"name": "tdd-london-swarm", "description": "Test-Driven Development with London school swarm", "keywords": ["TDD", "test-driven", "London", "mocking"], "category": "testing", "example_prompts": ["TDD development", "Write tests first", "Mock-based testing"]}
  ],
  "mcp_tools": [
    {"name": "swarm_init", "description": "Initialize multi-agent swarm with topology configuration", "keywords": ["swarm", "init", "topology", "coordination"], "category": "swarm", "example_prompts": ["Initialize swarm", "Set up agent coordination", "Configure topology"]},
    {"name": "agent_spawn", "description": "Spawn a new agent with intelligent model selection", "keywords": ["agent", "spawn", "create", "model"], "category": "agent", "example_prompts": ["Spawn coder agent", "Create new agent", "Add agent to swarm"]},
    {"name": "agent_terminate", "description": "Terminate an active agent", "keywords": ["agent", "terminate", "stop", "kill"], "category": "agent", "example_prompts": ["Stop agent", "Terminate worker", "Kill agent process"]},
    {"name": "agent_status", "description": "Get current status of an agent", "keywords": ["agent", "status", "health", "info"], "category": "agent", "example_prompts": ["Check agent status", "Get agent info", "Agent health check"]},
    {"name": "agent_list", "description": "List all agents with optional filtering", "keywords": ["agent", "list", "filter", "query"], "category": "agent", "example_prompts": ["List all agents", "Show active agents", "Filter agents by type"]},
    {"name": "memory_store", "description": "Store a value in persistent memory", "keywords": ["memory", "store", "save", "persist"], "category": "memory", "example_prompts": ["Store in memory", "Save pattern", "Persist data"]},
    {"name": "memory_retrieve", "description": "Retrieve a value from memory", "keywords": ["memory", "retrieve", "get", "load"], "category": "memory", "example_prompts": ["Get from memory", "Retrieve pattern", "Load stored data"]},
    {"name": "memory_search", "description": "Semantic vector search in memory", "keywords": ["memory", "search", "semantic", "vector"], "category": "memory", "example_prompts": ["Search memory", "Find similar patterns", "Semantic search"]},
    {"name": "task_create", "description": "Create a new task with priority and assignment", "keywords": ["task", "create", "assign", "priority"], "category": "task", "example_prompts": ["Create task", "Add new task", "Assign work"]},
    {"name": "task_status", "description": "Get task status and progress", "keywords": ["task", "status", "progress", "tracking"], "category": "task", "example_prompts": ["Check task status", "Get progress", "Track task"]},
    {"name": "hooks_pre-task", "description": "Record task start and get agent suggestions with intelligent model routing", "keywords": ["hooks", "pre-task", "routing", "suggestions"], "category": "hooks", "example_prompts": ["Pre-task coordination", "Get routing suggestion", "Start task hook"]},
    {"name": "hooks_post-task", "description": "Record task completion for learning", "keywords": ["hooks", "post-task", "learning", "completion"], "category": "hooks", "example_prompts": ["Post-task learning", "Record completion", "Train on result"]},
    {"name": "hooks_intelligence", "description": "RuVector intelligence system with SONA, MoE, HNSW", "keywords": ["intelligence", "SONA", "MoE", "HNSW", "neural"], "category": "intelligence", "example_prompts": ["Enable intelligence", "Check neural status", "SONA adaptation"]},
    {"name": "hooks_worker-dispatch", "description": "Dispatch background worker for analysis/optimization", "keywords": ["worker", "dispatch", "background", "async"], "category": "workers", "example_prompts": ["Dispatch audit worker", "Run optimization", "Background analysis"]},
    {"name": "neural_train", "description": "Train a neural model on patterns", "keywords": ["neural", "train", "model", "learning"], "category": "neural", "example_prompts": ["Train neural model", "Learn patterns", "Model training"]},
    {"name": "neural_predict", "description": "Make predictions using neural model", "keywords": ["neural", "predict", "inference", "model"], "category": "neural", "example_prompts": ["Predict action", "Neural inference", "Get prediction"]},
    {"name": "performance_benchmark", "description": "Run performance benchmarks", "keywords": ["performance", "benchmark", "metrics", "speed"], "category": "performance", "example_prompts": ["Run benchmarks", "Measure performance", "Speed test"]},
    {"name": "performance_bottleneck", "description": "Detect performance bottlenecks", "keywords": ["performance", "bottleneck", "analysis", "optimization"], "category": "performance", "example_prompts": ["Find bottlenecks", "Performance analysis", "Detect slowdowns"]},
    {"name": "github_repo_analyze", "description": "Analyze a GitHub repository", "keywords": ["GitHub", "repository", "analysis", "code"], "category": "github", "example_prompts": ["Analyze repo", "GitHub analysis", "Repository scan"]},
    {"name": "github_pr_manage", "description": "Manage pull requests", "keywords": ["GitHub", "PR", "pull-request", "manage"], "category": "github", "example_prompts": ["Manage PR", "Create pull request", "PR operations"]},
    {"name": "hive-mind_init", "description": "Initialize hive-mind collective", "keywords": ["hive-mind", "init", "collective", "coordination"], "category": "consensus", "example_prompts": ["Initialize hive", "Start collective", "Hive-mind setup"]},
    {"name": "hive-mind_consensus", "description": "Propose or vote on consensus", "keywords": ["hive-mind", "consensus", "vote", "proposal"], "category": "consensus", "example_prompts": ["Propose consensus", "Vote on decision", "Collective agreement"]},
    {"name": "embeddings_generate", "description": "Generate embeddings for text", "keywords": ["embeddings", "generate", "vector", "text"], "category": "embeddings", "example_prompts": ["Generate embedding", "Text to vector", "Create embedding"]},
    {"name": "embeddings_search", "description": "Semantic search across stored embeddings", "keywords": ["embeddings", "search", "semantic", "similarity"], "category": "embeddings", "example_prompts": ["Search embeddings", "Semantic search", "Find similar"]},
    {"name": "aidefence_scan", "description": "Scan input for AI manipulation threats", "keywords": ["security", "scan", "threats", "injection"], "category": "security", "example_prompts": ["Scan for threats", "Security check", "Detect injection"]},
    {"name": "claims_claim", "description": "Claim an issue for work", "keywords": ["claims", "issue", "work", "assignment"], "category": "claims", "example_prompts": ["Claim issue", "Take work item", "Assign to self"]},
    {"name": "workflow_create", "description": "Create a new workflow", "keywords": ["workflow", "create", "automation", "process"], "category": "workflow", "example_prompts": ["Create workflow", "Define process", "Automation setup"]},
    {"name": "workflow_execute", "description": "Execute a workflow", "keywords": ["workflow", "execute", "run", "automation"], "category": "workflow", "example_prompts": ["Run workflow", "Execute process", "Start automation"]},
    {"name": "session_save", "description": "Save current session state", "keywords": ["session", "save", "state", "persist"], "category": "session", "example_prompts": ["Save session", "Persist state", "Store session"]},
    {"name": "session_restore", "description": "Restore a saved session", "keywords": ["session", "restore", "load", "recover"], "category": "session", "example_prompts": ["Restore session", "Load state", "Recover session"]},
    {"name": "system_status", "description": "Get overall system status", "keywords": ["system", "status", "health", "overview"], "category": "system", "example_prompts": ["System status", "Health check", "System overview"]},
    {"name": "coordination_orchestrate", "description": "Orchestrate multi-agent coordination", "keywords": ["coordination", "orchestrate", "multi-agent", "parallel"], "category": "coordination", "example_prompts": ["Orchestrate agents", "Coordinate task", "Parallel execution"]}
  ],
  "agentdb_cli": [
    {"name": "agentdb init", "description": "Initialize database with schema and configuration", "keywords": ["init", "setup", "database", "schema"], "category": "database", "example_prompts": ["Initialize AgentDB", "Setup vector database", "Create database schema"]},
    {"name": "agentdb query", "description": "Query vectors with similarity search", "keywords": ["query", "search", "vector", "similarity"], "category": "search", "example_prompts": ["Query vectors", "Search database", "Find similar vectors"]},
    {"name": "agentdb pattern store", "description": "Store reasoning patterns (388K ops/sec)", "keywords": ["pattern", "store", "save", "reasoning"], "category": "patterns", "example_prompts": ["Store pattern", "Save reasoning", "Add to pattern library"]},
    {"name": "agentdb pattern search", "description": "Semantic pattern retrieval (32.6M ops/sec)", "keywords": ["pattern", "search", "semantic", "retrieval"], "category": "patterns", "example_prompts": ["Search patterns", "Find similar patterns", "Pattern retrieval"]},
    {"name": "agentdb reflexion store", "description": "Store episodic learning experience", "keywords": ["reflexion", "episode", "learning", "experience"], "category": "learning", "example_prompts": ["Store episode", "Save experience", "Record learning"]},
    {"name": "agentdb reflexion retrieve", "description": "Retrieve similar episodes", "keywords": ["reflexion", "retrieve", "episodes", "similar"], "category": "learning", "example_prompts": ["Get episodes", "Find similar experiences", "Retrieve learning"]},
    {"name": "agentdb skill create", "description": "Create reusable skill (304 ops/sec)", "keywords": ["skill", "create", "reusable", "code"], "category": "skills", "example_prompts": ["Create skill", "Define reusable function", "Add skill"]},
    {"name": "agentdb skill search", "description": "Discover applicable skills (694 ops/sec)", "keywords": ["skill", "search", "discover", "match"], "category": "skills", "example_prompts": ["Search skills", "Find applicable skill", "Discover skills"]},
    {"name": "agentdb skill consolidate", "description": "Auto-extract skills from episodes", "keywords": ["skill", "consolidate", "extract", "automatic"], "category": "skills", "example_prompts": ["Consolidate skills", "Extract from episodes", "Auto-generate skills"]},
    {"name": "agentdb learner run", "description": "Discover causal patterns", "keywords": ["learner", "causal", "patterns", "discovery"], "category": "learning", "example_prompts": ["Run learner", "Discover patterns", "Causal analysis"]},
    {"name": "agentdb simulate", "description": "Run latent space simulations (25 scenarios)", "keywords": ["simulate", "latent-space", "scenarios", "testing"], "category": "simulation", "example_prompts": ["Run simulation", "Test scenarios", "Latent space analysis"]},
    {"name": "agentdb benchmark", "description": "Run comprehensive performance benchmarks", "keywords": ["benchmark", "performance", "speed", "testing"], "category": "performance", "example_prompts": ["Run benchmarks", "Test performance", "Measure speed"]},
    {"name": "agentdb prune", "description": "Intelligent data cleanup", "keywords": ["prune", "cleanup", "optimization", "storage"], "category": "maintenance", "example_prompts": ["Prune database", "Clean old data", "Optimize storage"]},
    {"name": "agentdb stats", "description": "Get database statistics (8.8x faster cached)", "keywords": ["stats", "statistics", "metrics", "info"], "category": "monitoring", "example_prompts": ["Get stats", "Database metrics", "Show statistics"]},
    {"name": "agentdb create-plugin", "description": "Create learning plugin from template", "keywords": ["plugin", "create", "template", "learning"], "category": "plugins", "example_prompts": ["Create plugin", "Generate from template", "New learning plugin"]},
    {"name": "agentdb mcp", "description": "Start MCP server for Claude Code integration", "keywords": ["mcp", "server", "Claude", "integration"], "category": "integration", "example_prompts": ["Start MCP server", "Claude integration", "Enable MCP tools"]},
    {"name": "agentdb export", "description": "Export database to JSON", "keywords": ["export", "backup", "JSON", "data"], "category": "data", "example_prompts": ["Export database", "Backup data", "Save to JSON"]},
    {"name": "agentdb import", "description": "Import data from JSON", "keywords": ["import", "restore", "JSON", "data"], "category": "data", "example_prompts": ["Import data", "Restore backup", "Load from JSON"]}
  ],
  "background_workers": [
    {"name": "ultralearn", "description": "Deep knowledge acquisition worker", "priority": "normal", "keywords": ["learning", "knowledge", "deep", "acquisition"], "example_prompts": ["Deep learning analysis", "Acquire knowledge", "Learn from codebase"]},
    {"name": "optimize", "description": "Performance optimization worker", "priority": "high", "keywords": ["optimize", "performance", "speed", "efficiency"], "example_prompts": ["Optimize performance", "Improve speed", "Efficiency analysis"]},
    {"name": "consolidate", "description": "Memory consolidation worker", "priority": "low", "keywords": ["consolidate", "memory", "merge", "cleanup"], "example_prompts": ["Consolidate memory", "Merge patterns", "Memory cleanup"]},
    {"name": "predict", "description": "Predictive preloading worker", "priority": "normal", "keywords": ["predict", "preload", "anticipate", "cache"], "example_prompts": ["Predict needs", "Preload resources", "Anticipate requests"]},
    {"name": "audit", "description": "Security analysis worker", "priority": "critical", "keywords": ["audit", "security", "analysis", "vulnerabilities"], "example_prompts": ["Security audit", "Find vulnerabilities", "Scan for issues"]},
    {"name": "map", "description": "Codebase mapping worker", "priority": "normal", "keywords": ["map", "codebase", "structure", "analysis"], "example_prompts": ["Map codebase", "Analyze structure", "Create code map"]},
    {"name": "preload", "description": "Resource preloading worker", "priority": "low", "keywords": ["preload", "resources", "cache", "prefetch"], "example_prompts": ["Preload resources", "Cache data", "Prefetch files"]},
    {"name": "deepdive", "description": "Deep code analysis worker", "priority": "normal", "keywords": ["deepdive", "analysis", "code", "detailed"], "example_prompts": ["Deep code analysis", "Detailed investigation", "Thorough review"]},
    {"name": "document", "description": "Auto-documentation worker", "priority": "normal", "keywords": ["document", "documentation", "auto", "generate"], "example_prompts": ["Auto-document code", "Generate docs", "Create documentation"]},
    {"name": "refactor", "description": "Refactoring suggestions worker", "priority": "normal", "keywords": ["refactor", "suggestions", "improve", "clean"], "example_prompts": ["Suggest refactoring", "Improve code", "Clean up codebase"]},
    {"name": "benchmark", "description": "Performance benchmarking worker", "priority": "normal", "keywords": ["benchmark", "performance", "measure", "metrics"], "example_prompts": ["Run benchmarks", "Measure performance", "Get metrics"]},
    {"name": "testgaps", "description": "Test coverage analysis worker", "priority": "normal", "keywords": ["testgaps", "coverage", "tests", "missing"], "example_prompts": ["Find test gaps", "Coverage analysis", "Missing tests"]}
  ],
  "performance_metrics": {
    "flash_attention_speedup": "2.49x-7.47x",
    "memory_reduction": "50-75%",
    "hnsw_search_improvement": "150x-12,500x",
    "pattern_search_ops_per_sec": "32.6M",
    "pattern_store_ops_per_sec": "388K",
    "batch_insert_improvement": "500x",
    "vector_search_latency": "<100us",
    "pattern_retrieval_latency": "<1ms",
    "sona_adaptation_latency": "<0.05ms",
    "mcp_response_target": "<100ms",
    "cli_startup_target": "<500ms",
    "agent_booster_speedup": "352x",
    "gnn_recall_improvement": "+12.4%"
  },
  "integration_ecosystem": [
    {"name": "agentdb", "description": "High-performance vector database with HNSW indexing", "package": "agentdb@alpha"},
    {"name": "ruv-swarm", "description": "Multi-agent swarm coordination", "package": "ruv-swarm"},
    {"name": "flow-nexus", "description": "Workflow automation and nexus", "package": "flow-nexus@latest"},
    {"name": "ruvector", "description": "Rust-based vector operations with SIMD", "package": "ruvector"},
    {"name": "@ruvector/core", "description": "Core RuVector functionality", "package": "@ruvector/core"},
    {"name": "@ruvector/router", "description": "Intelligent routing system", "package": "@ruvector/router"},
    {"name": "@ruvector/ruvllm", "description": "RuvLLM local inference", "package": "@ruvector/ruvllm"},
    {"name": "@ruvector/sona", "description": "Self-Optimizing Neural Architecture", "package": "@ruvector/sona"},
    {"name": "@ruvector/attention", "description": "Attention mechanisms", "package": "@ruvector/attention"},
    {"name": "@ruvector/tiny-dancer", "description": "Lightweight neural inference", "package": "@ruvector/tiny-dancer"},
    {"name": "fastmcp", "description": "Fast MCP server implementation", "package": "fastmcp"},
    {"name": "@anthropic-ai/claude-agent-sdk", "description": "Claude Agent SDK", "package": "@anthropic-ai/claude-agent-sdk"}
  ],
  "attention_mechanisms": [
    {"name": "Flash Attention", "description": "Memory-efficient attention with 2.49x-7.47x speedup and 50-75% memory reduction", "keywords": ["flash", "attention", "memory-efficient", "speedup"]},
    {"name": "Multi-Head Attention", "description": "8-head attention configuration for parallel processing", "keywords": ["multi-head", "attention", "parallel", "heads"]},
    {"name": "Linear Attention", "description": "O(n) complexity for long sequences", "keywords": ["linear", "attention", "complexity", "sequences"]},
    {"name": "Hyperbolic Attention", "description": "For hierarchical structures using Poincare ball", "keywords": ["hyperbolic", "attention", "hierarchical", "poincare"]},
    {"name": "MoE Attention", "description": "Mixture of Experts routing for specialized attention", "keywords": ["MoE", "attention", "experts", "routing"]},
    {"name": "GraphRoPE", "description": "Topology-aware position embeddings", "keywords": ["graph", "RoPE", "topology", "position"]}
  ],
  "learning_algorithms": [
    {"name": "Decision Transformer", "description": "Sequence modeling RL for offline learning from logged experiences", "keywords": ["decision-transformer", "offline-RL", "sequence", "imitation"]},
    {"name": "Q-Learning", "description": "Value-based off-policy learning for discrete actions", "keywords": ["q-learning", "value-based", "discrete", "off-policy"]},
    {"name": "SARSA", "description": "On-policy TD learning for safe exploration", "keywords": ["sarsa", "on-policy", "TD", "safe"]},
    {"name": "Actor-Critic", "description": "Policy gradient with value baseline for continuous control", "keywords": ["actor-critic", "policy-gradient", "continuous", "baseline"]},
    {"name": "Active Learning", "description": "Query-based learning for label efficiency", "keywords": ["active-learning", "query", "labels", "uncertainty"]},
    {"name": "Adversarial Training", "description": "Robustness enhancement against perturbations", "keywords": ["adversarial", "training", "robustness", "defense"]},
    {"name": "Curriculum Learning", "description": "Progressive difficulty training for complex tasks", "keywords": ["curriculum", "progressive", "difficulty", "training"]},
    {"name": "Federated Learning", "description": "Privacy-preserving distributed learning", "keywords": ["federated", "distributed", "privacy", "collaborative"]},
    {"name": "Multi-Task Learning", "description": "Transfer learning across related tasks", "keywords": ["multi-task", "transfer", "knowledge", "sharing"]}
  ],
  "consensus_protocols": [
    {"name": "Byzantine", "description": "BFT consensus tolerating f < n/3 faulty nodes", "keywords": ["byzantine", "BFT", "fault-tolerant", "consensus"]},
    {"name": "Raft", "description": "Leader-based consensus tolerating f < n/2 failures", "keywords": ["raft", "leader", "election", "log-replication"]},
    {"name": "Gossip", "description": "Epidemic protocol for eventual consistency", "keywords": ["gossip", "epidemic", "eventual", "consistency"]},
    {"name": "CRDT", "description": "Conflict-free replicated data types", "keywords": ["CRDT", "conflict-free", "replicated", "distributed"]},
    {"name": "Quorum", "description": "Configurable quorum-based consensus", "keywords": ["quorum", "configurable", "majority", "consensus"]}
  ],
  "topologies": [
    {"name": "hierarchical", "description": "Queen controls workers directly (anti-drift for small teams)", "keywords": ["hierarchical", "queen", "workers", "control"]},
    {"name": "hierarchical-mesh", "description": "V3 queen + peer communication (recommended for 10+ agents)", "keywords": ["hierarchical-mesh", "hybrid", "peer", "queen"]},
    {"name": "mesh", "description": "Fully connected peer network", "keywords": ["mesh", "peer", "connected", "distributed"]},
    {"name": "ring", "description": "Circular communication pattern", "keywords": ["ring", "circular", "sequential", "communication"]},
    {"name": "star", "description": "Central coordinator with spokes", "keywords": ["star", "central", "coordinator", "spokes"]},
    {"name": "adaptive", "description": "Dynamic topology switching based on load", "keywords": ["adaptive", "dynamic", "switching", "automatic"]}
  ],
  "quantization_types": [
    {"name": "binary", "description": "32x memory reduction, 10x faster, 95-98% accuracy", "keywords": ["binary", "quantization", "compression", "fast"]},
    {"name": "scalar", "description": "4x memory reduction, 3x faster, 98-99% accuracy", "keywords": ["scalar", "quantization", "balanced", "efficient"]},
    {"name": "product", "description": "8-16x memory reduction, 5x faster, 93-97% accuracy", "keywords": ["product", "quantization", "compression", "high-dim"]},
    {"name": "none", "description": "Full precision, maximum accuracy", "keywords": ["none", "full-precision", "accurate", "uncompressed"]}
  ]
}