Tags

Tags give the ability to mark specific points in history as being important
  • v3.0.0

    Universal Agent Architecture v3.0.0
    
    Major architectural change implementing Anthropic's recommendation:
    "Don't Build Agents, Build Skills Instead"
    
    Changes:
    - 88 agent classes → 1 UniversalAgent
    - Dynamic skill loading (106 skills)
    - Parallel execution preserved via WorkerPool
    - Circuit breaker for Human-in-the-Loop safety
    - Checkpoint/resume for recovery
    - 9 core skills created
    - 30-50% token reduction expected
    
    New API endpoints:
    - POST /agents/universal/invoke
    - POST /agents/oneshot/invoke
    
    Breaking: Old agent classes still work but Universal Agent is now preferred.
  • v2.5.0

    v2.5.0: Agent Autocomplete with Full List
  • v2.4.0

    v2.4.0: Self-Improvement Loop - Phase 8 Adaptive Learning System
    
    Complete Adaptive Learning implementation:
    - Feedback Collection (explicit, implicit, outcomes)
    - Learning Engine (preferences, patterns, corrections)
    - Adaptation Manager (behavior modification, regression prevention)
    - Performance Tracking & Quality Analysis
    
    Target metrics: 89% acceptance rate, 1.1 corrections/task, 4.7/5.0 satisfaction
  • v2.2.1

    Release v2.2.1: Enhanced Agent and Plugin System Documentation
    
    - Documented all 88+ agents across 12 categories
    - Documented all 88+ plugins across 15 categories
    - Enhanced Agent System section with voice families and interaction patterns
    - Enhanced Plugin System section with usage patterns and workflows
    - Version bump: 2.2.0 → 2.2.1 (PATCH - documentation only)
    - Published to npm: @drtonylove1963/[email protected]
  • v2.1.0

    Release v2.1.0 - Phase II.4 Complete
    
    ✅ 100% Test Coverage (323/323 tests passing)
    ✅ Code Quality: 9.6/10
    ✅ Production Ready
    
    - Fixed 71 deprecation warnings (datetime.utcnow → datetime.now(UTC))
    - GitHub Marketplace Integration (GitHubAPIClient)
    - AI-Powered Plugin Generation (enhanced)
    - TokenMonitor Integration (real-time tracking)
    - Phase III Architecture Planning (26 weeks, $725K)
    
    - Phase II.1 (AI Enhancement): 11/11 tests
    - Phase II.2 (Enterprise/Platform): 70/70 tests
    - Phase II.3 (Cloud/Scale): 80/80 tests
    - Phase II.4 (Completion): 88 new tests
    - Total: 323/323 tests passing
    
    - PHASE_III_ARCHITECTURE.md (20+ pages)
    - PHASE_III_CHECKLIST.md (150+ tasks)
    - PHASE_III_SUMMARY.md
    - PHASE_III_DIAGRAMS.md (15 diagrams)
    
    See CHANGELOG.md and PROJECT_STATUS_REPORT.md for details.
  • v2.0.0

    Release v2.0.0: Multi-Agent Context Management
    
    Major Features:
    - Multi-agent system with 5 specialized roles
    - Knowledge extraction engine (60-70% token reduction)
    - Agent context management and switching
    - Structured knowledge graphs
    - Entity and relationship extraction
    - Migration tools from v1.x
    
    Full release notes in CHANGELOG.md
  • v1.0.0

    Release v1.0.0: Initial release with full CLI functionality