|
165 | 165 | - ✅ **Smart Trace Validation**: Documented and implemented in `/analyze` command (lines 209-236) |
166 | 166 | - ⚠️ **Task-to-Issues Command**: Template exists at `templates/commands/taskstoissues.md` but only supports GitHub (multi-tracker support planned in Tier 1.5) |
167 | 167 |
|
| 168 | +### **Lean Architecture Improvements** *(100% Complete)* - **COMPLETED** - 6 Architecture Enhancements |
| 169 | + |
| 170 | +Major enhancements to the architecture system making it more flexible, intelligent, and integrated with existing documentation. |
| 171 | + |
| 172 | +- ✅ **Lean Architecture Views**: Configurable view generation (core 5 + optional 2) |
| 173 | + - Default "core" views: Context, Functional, Information, Development, Deployment |
| 174 | + - Optional views: Concurrency, Operational (via `--views` flag) |
| 175 | + - `--views all`, `--views core`, `--views concurrency,operational` |
| 176 | + |
| 177 | +- ✅ **Surprise-Value Heuristic**: Skip obvious ecosystem defaults, document only surprising/risky decisions |
| 178 | + - `--adr-heuristic surprising` (default), `all`, `minimal` |
| 179 | + - Configuration in `config.json`: `architecture.adr.heuristic` |
| 180 | + |
| 181 | +- ✅ **Constitution Cross-Reference**: Strict checking for ADR/Constitution alignment |
| 182 | + - Always enabled in `/architect.clarify` |
| 183 | + - Detects duplicates, violations, unclear alignment |
| 184 | + - **Option A (Amend Constitution) as PRIMARY resolution** |
| 185 | + |
| 186 | +- ✅ **ADR Template Improvements**: |
| 187 | + - "Common Alternatives" (not "Alternatives Considered") |
| 188 | + - Neutral "Trade-offs" framing (not "Rejected because") |
| 189 | + - "Discovered" status for reverse-engineered ADRs |
| 190 | + - "Constitution Alignment" section |
| 191 | + |
| 192 | +- ✅ **Existing Docs Deduplication**: Scan and reference instead of duplicate |
| 193 | + - Scans `docs/` directory and root `*.md` files |
| 194 | + - References existing docs (README, AGENTS.md, CONTRIBUTING) |
| 195 | + - Auto-merges when existing architecture found |
| 196 | + |
| 197 | +- ✅ **Risks & Gaps Analysis**: Cross-cutting analysis in `/architect.clarify` |
| 198 | + - Identifies operational gaps, tech debt, SPOFs, security concerns |
| 199 | + - Section-based gap IDs (e.g., `3.6.1`) |
| 200 | + - Runs BEFORE constitution cross-reference |
| 201 | + |
| 202 | +**Files Modified**: 13 files including templates, scripts, and configuration |
| 203 | + |
| 204 | +--- |
| 205 | + |
168 | 206 | ### **Optional Architecture Support** *(100% Complete)* - **COMPLETED** - Enterprise Architecture Features |
169 | 207 |
|
170 | 208 | Architecture support is now available in all workflow modes as optional commands. The `/architect` and `/constitution` commands work silently in any mode, with no warnings if files are missing. |
@@ -742,6 +780,115 @@ Consolidates: Command-level model selection + context budgeting + two-model revi |
742 | 780 |
|
743 | 781 | ### **Future Enhancement Categories** |
744 | 782 |
|
| 783 | +### **Compound Engineering Integration** *(0% Complete)* - **MEDIUM PRIORITY** - Insights from EveryInc's Compound Engineering Plugin |
| 784 | + |
| 785 | +**Reference**: Analysis of [EveryInc/compound-engineering-plugin](https://github.com/EveryInc/compound-engineering-plugin) (7.5k stars, 598 forks) |
| 786 | + |
| 787 | +**Key Insight**: Compound Engineering Plugin focuses on workflow efficiency through multi-agent parallelization and knowledge compounding, while Agentic SDLC Spec Kit focuses on structured SDLC methodology. These approaches are complementary. |
| 788 | + |
| 789 | +#### **Swarm Mode Parallel Execution** *(0% Complete)* - **HIGH PRIORITY** |
| 790 | + |
| 791 | +- **Description**: Implement swarm mode execution (like Compound's `/slfg`) that runs multiple sub-agents in parallel for tasks marked with `[P]` markers, dramatically accelerating development workflows |
| 792 | +- **Key Components**: |
| 793 | + - Parallel agent spawning for independent tasks within phases |
| 794 | + - Worktree-based isolation (validates incident.io pattern with 5+ agents) |
| 795 | + - Coordinated execution with dependency awareness |
| 796 | + - Phase-respecting orchestration (Setup → Foundational → Stories) |
| 797 | +- **Benefits**: |
| 798 | + - 3-5x faster execution for parallelizable tasks |
| 799 | + - Validates incident.io multi-agent + worktree pattern |
| 800 | + - Maintains code isolation and merge safety |
| 801 | +- **Implementation**: Extend existing `[P]` marker system with swarm orchestration layer |
| 802 | +- **References**: incident.io workflow pattern, Worktrunk integration |
| 803 | + |
| 804 | +#### **Multi-Agent Review System** *(0% Complete)* - **MEDIUM PRIORITY** |
| 805 | + |
| 806 | +- **Description**: Implement specialized reviewer agents (like Compound's 15 review agents) that perform focused code reviews: |
| 807 | + - `security-sentinel`: Security vulnerability detection |
| 808 | + - `performance-oracle`: Performance optimization suggestions |
| 809 | + - `rails-reviewer`: Framework-specific best practices |
| 810 | + - `accessibility-guardian`: A11y compliance checking |
| 811 | + - `test-coverage-analyst`: Test quality validation |
| 812 | +- **Key Components**: |
| 813 | + - Temperature-tuned agents (review: 0.1, planning: 0.2, creative: 0.6) |
| 814 | + - Parallel review execution before merge |
| 815 | + - Review agent selection based on code changes |
| 816 | + - Aggregated review reports with actionable items |
| 817 | +- **Benefits**: |
| 818 | + - Comprehensive pre-merge validation |
| 819 | + - Specialized expertise without context bloat |
| 820 | + - Consistent quality gates across team |
| 821 | +- **Integration**: Hook into `/spec.analyze` and pre-merge workflows |
| 822 | + |
| 823 | +#### **Knowledge Compounding System** *(0% Complete)* - **MEDIUM PRIORITY** |
| 824 | + |
| 825 | +- **Description**: Enhance `/spec.levelup` with systematic knowledge capture inspired by Compound's philosophy: "Each unit of engineering work should make subsequent units easier" |
| 826 | +- **Key Components**: |
| 827 | + - Pattern extraction from completed features |
| 828 | + - Anti-pattern documentation from failed approaches |
| 829 | + - Reusable solution templates for common problems |
| 830 | + - Team knowledge base with searchable patterns |
| 831 | + - Automatic skill generation from recurring patterns |
| 832 | +- **Workflow Enhancement**: |
| 833 | + ``` |
| 834 | + Plan → Work → Review → Compound → Repeat |
| 835 | + ``` |
| 836 | +- **Benefits**: |
| 837 | + - Institutional knowledge preservation |
| 838 | + - Accelerated future development through pattern reuse |
| 839 | + - Reduced repeated mistakes |
| 840 | + - Team learning velocity increase |
| 841 | +- **Integration**: Extend existing `/spec.levelup` with structured knowledge packets |
| 842 | + |
| 843 | +#### **Browser Automation Integration** *(0% Complete)* - **LOW PRIORITY** |
| 844 | + |
| 845 | +- **Description**: Integrate browser automation capabilities (like Compound's agent-browser CLI) for e2e testing and UI validation |
| 846 | +- **Key Components**: |
| 847 | + - Automated e2e test generation from specifications |
| 848 | + - Visual regression testing integration |
| 849 | + - UI interaction recording and replay |
| 850 | + - Screenshot-based validation in `/spec.implement` |
| 851 | +- **Benefits**: |
| 852 | + - Automated visual validation |
| 853 | + - Reduced manual QA effort |
| 854 | + - Better spec-to-implementation alignment |
| 855 | +- **Implementation**: Integration with Playwright, Puppeteer, or agent-browser |
| 856 | + |
| 857 | +#### **Cross-Platform Plugin Converter** *(0% Complete)* - **LOW PRIORITY** |
| 858 | + |
| 859 | +- **Description**: Create plugin conversion system (like Compound's Bun/TypeScript CLI) to convert between agent formats: |
| 860 | + - Claude Code plugins ↔ OpenCode format |
| 861 | + - Claude Code plugins ↔ Codex format |
| 862 | + - Custom agent format support |
| 863 | +- **Key Components**: |
| 864 | + - Agent format parsers and transformers |
| 865 | + - Tool mapping between platforms (bash/read/write/edit) |
| 866 | + - Temperature inference from agent names |
| 867 | + - MCP server conversion |
| 868 | +- **Benefits**: |
| 869 | + - Broader agent ecosystem compatibility |
| 870 | + - Reduced vendor lock-in |
| 871 | + - Easier migration between AI tools |
| 872 | +- **Implementation**: Python-based converter leveraging existing AGENT_CONFIG |
| 873 | + |
| 874 | +#### **Task Delegation Templates** *(0% Complete)* - **MEDIUM PRIORITY* |
| 875 | + |
| 876 | +- **Description**: Create rich task delegation system with specialized agent types inspired by Compound's agent ecosystem: |
| 877 | + - **Explore Agents**: Read-only codebase search (fresh slate, no context inheritance) |
| 878 | + - **Plan Agents**: Design/implementation planning (full context inheritance) |
| 879 | + - **Execute Agents**: Multi-step task execution with tool sequencing |
| 880 | + - **Review Agents**: Focused quality validation |
| 881 | +- **Key Components**: |
| 882 | + - Agent type selection matrix based on task characteristics |
| 883 | + - Context inheritance rules per agent type |
| 884 | + - Model selection guidance per agent type |
| 885 | + - "When NOT to Spawn" guidelines to prevent context bloat |
| 886 | +- **Benefits**: |
| 887 | + - Optimized agent utilization |
| 888 | + - Reduced unnecessary context window usage |
| 889 | + - Better task-agent matching |
| 890 | +- **Integration**: Extend Tier 2 Sub-Agent Coordination Framework |
| 891 | + |
745 | 892 | ### **Beads-Backed Task Execution Tracker** *(0% Complete)* - **MEDIUM PRIORITY** - Agent-native persistent work tracking |
746 | 893 |
|
747 | 894 | **Problem Statement**: |
|
0 commit comments