@mc1global/opencode-jarvis 0.12.0 → 0.13.1

package/.opencode/devlogs/BrainWaves-iPad-Support-continued/2026-02-09-1938-jarvis-opencode.md

@@ -0,0 +1,34 @@
+# Devlog: jarvis-opencode
+
+**Devlog ID**: 2026-02-09-1938-BrainWaves-iPad-Support-continued
+**Instance**: BrainWaves-iPad-Support-continued
+**Started**: 2026-02-09 19:38
+**Workspace**: jarvis-opencode
+
+---
+
+## Overview
+
+(Describe the goals and scope of this development log)
+
+---
+
+## Progress
+
+(Progress updates will be added here)
+
+---
+
+## Notes
+
+(Additional notes)
+
+---
+
+---
+
+## Devlog Closed
+
+**Ended**: 2026-02-09 19:38
+
+(Add comprehensive summary here)

package/.opencode/devlogs/EP-005 Sprint 6: SSE Streaming Implementation/.current-devlog

@@ -0,0 +1 @@
+2026-02-09-2332-neural-agents.md

package/.opencode/devlogs/EP-005 Sprint 6: SSE Streaming Implementation/2026-02-09-2332-neural-agents.md

@@ -0,0 +1,26 @@
+# Devlog: neural-agents
+
+**Devlog ID**: 2026-02-09-2332-EP-005 Sprint 6: SSE Streaming Implementation
+**Instance**: EP-005 Sprint 6: SSE Streaming Implementation
+**Started**: 2026-02-09 23:32
+**Workspace**: jarvis-opencode
+
+---
+
+## Overview
+
+(Describe the goals and scope of this development log)
+
+---
+
+## Progress
+
+(Progress updates will be added here)
+
+---
+
+## Notes
+
+(Additional notes)
+
+---

package/.opencode/devlogs/JARVIS/2026-01-14-2315-agents-skills-review.md

@@ -0,0 +1,270 @@
+# Session: Agents & Skills Review
+**Started**: 2026-01-14 23:15
+
+## Overview
+Systematic review of all .claude agents and skills for conformity with official Claude Code CLI specification.
+
+## Goals
+1. Review each agent and skill for conformity with official spec
+2. Identify non-conforming elements (missing fields, incorrect structure)
+3. Fix issues found in each entity
+4. Document findings and corrections
+
+## Scope
+
+### Agents (7)
+| # | Agent | Status |
+|---|-------|--------|
+| 1 | code-reviewer.md | **DONE** (455→100, -78%) |
+| 2 | data-engineer.md | **DONE** (230→100, -57%) |
+| 3 | jarvis.md | **DONE** (195→105, -46%) |
+| 4 | product-owner.md | **DONE** (314→103, -67%) |
+| 5 | script-developer.md | **DONE** (364→122, -66%) |
+| 6 | script-developer-optimized.md | **DONE** (330→105, -68%) |
+| 7 | solution-architect.md | **DONE** (354→111, -69%) |
+
+### Skills (43)
+| # | Skill | Status |
+|---|-------|--------|
+| 1 | bash-optimizer | pending |
+| 2 | boilerplate-sync | pending |
+| 3 | code-block-manager | pending |
+| 4 | code-intelligence | pending |
+| 5 | code-validator | pending |
+| 6 | command-intelligence | pending |
+| 7 | component-generator | pending |
+| 8 | config-manager | pending |
+| 9 | context-memory-guide | pending |
+| 10 | csv-manager | pending |
+| 11 | data-validator | pending |
+| 12 | doc-generator | pending |
+| 13 | env-manager | pending |
+| 14 | file-manager | pending |
+| 15 | git-repo-manager | pending |
+| 16 | jarvis-config | pending |
+| 17 | jarvis-installer | pending |
+| 18 | json-manager | pending |
+| 19 | llm-manager | pending |
+| 20 | lsp-navigator | pending |
+| 21 | markdown-manager | pending |
+| 22 | migration-manager | pending |
+| 23 | mindfulness | pending |
+| 24 | model-router | pending |
+| 25 | ollama-model-finder | pending |
+| 26 | policy-manager | pending |
+| 27 | process-manager | pending |
+| 28 | quick-reference | pending |
+| 29 | rag-manager | pending |
+| 30 | reporting-framework | pending |
+| 31 | safe-script-development | pending |
+| 32 | schema-mapper | pending |
+| 33 | session-startup | pending |
+| 34 | snapshot-manager | pending |
+| 35 | task-complexity-analyzer | pending |
+| 36 | token-optimizer | pending |
+| 37 | toon-manager | pending |
+| 38 | user-story-chat | pending |
+| 39 | user-story-dev | pending |
+| 40 | user-story-manager | pending |
+| 41 | workspace-agents-mcp | pending |
+| 42 | workspace-manager | pending |
+| 43 | yaml-manager | pending |
+
+## Conformity Criteria
+
+### Agents
+- YAML frontmatter with `---` delimiters
+- Required: `name` (lowercase, hyphens), `description`
+- Optional: `tools` (comma-separated), `model` (sonnet/opus/haiku/inherit)
+- Body: System prompt in markdown
+
+### Skills
+- Folder with SKILL.md required
+- YAML frontmatter with `---` delimiters
+- Required: `name` (lowercase, numbers, hyphens, max 64 chars), `description` (max 1024 chars)
+- Optional: `allowed-tools`
+- Optional files: reference.md, examples.md, scripts/, templates/
+
+## Progress
+
+### Session Log
+
+#### 2026-01-14 23:28 - Phase 1 Complete (Proof-of-Concept)
+
+**Agent Reviewed**: code-reviewer.md
+**Status**: REFACTORED
+
+**Changes Made**:
+1. Created centralized script references structure:
+   - `claude_cli_resources/docs/scripts/context_memory/context_memory_cli.md` (59 lines)
+   - `claude_cli_resources/docs/scripts/context_memory/active_context.md` (57 lines)
+   - `claude_cli_resources/docs/scripts/rag/rag_cli.md` (75 lines)
+   - `claude_cli_resources/docs/scripts/code/lsp_navigator.md` (74 lines)
+   - `claude_cli_resources/docs/scripts/validation/policy_manager.md` (76 lines)
+
+2. Refactored code-reviewer.md:
+   - Before: 455 lines
+   - After: 100 lines
+   - Reduction: **78%**
+
+**Pattern Established**:
+- Agents reference tools via `@claude_cli_resources/docs/scripts/...`
+- Load on-demand: references read only when specific action needed
+- TOON format for compact structured data
+
+**Next**: Apply same pattern to remaining 6 agents
+
+**Validation**: Tested agent reviewing `scripts/rag/rag_cli.py` → APPROVED
+- Full review process executed successfully
+- All phases working (static analysis, architecture, security, governance)
+- Agent functionality preserved with 78% token reduction
+
+---
+
+#### 2026-01-14 23:35 - Starting Agent #2: data-engineer.md
+
+---
+
+#### 2026-01-14 23:50 - Phase 2 Complete (All Agents)
+
+**Summary**:
+- 7 agents refactored
+- Total: 2242 → 746 lines (-67%)
+- 6 script references created
+
+---
+
+#### 2026-01-14 23:51 - Starting Phase 3 (Skills)
+
+**Large Skills Refactored (4/43)**:
+- agent-invoker: 217 → 71 (-67%)
+- jarvis-config: 178 → 78 (-56%)
+- lsp-navigator: 100 → 58 (-42%)
+- rag-manager: 87 → 68 (-22%)
+
+**Remaining 39 skills**: Already compact (28-47 lines each)
+**Total skills**: 1802 → 1495 lines (-17%)
+
+---
+
+## FINAL SUMMARY - All Phases Complete
+
+### Phase 2: Agents (7)
+| Before | After | Savings |
+|--------|-------|---------|
+| 2242 lines | 746 lines | **-67%** |
+
+### Phase 3: Skills (43)
+| Before | After | Savings |
+|--------|-------|---------|
+| ~1802 lines | 1495 lines | **-17%** |
+
+### Script References Created (6 files)
+```
+claude_cli_resources/docs/scripts/
+├── context_memory/ (2 files)
+├── rag/ (1 file)
+├── code/ (1 file)
+├── validation/ (1 file)
+└── llm/ (1 file)
+```
+
+### Total Token Savings (Estimated)
+- **Agents**: ~22K → ~7.5K tokens (-66%)
+- **Skills (large)**: ~3K → ~1.5K tokens (-50%)
+- **Lazy loading benefit**: References loaded only when needed
+
+### Pattern Established
+1. Agents/Skills reference tools via `@claude_cli_resources/docs/scripts/...`
+2. TOON format for structured data
+3. Load on-demand: references read only when action needed
+
+---
+
+## SESSION END SUMMARY
+
+**Ended**: 2026-01-15
+**Duration**: ~2 hours (across context windows)
+
+### Git Summary
+
+**Files Changed**: 77 total
+- **Insertions**: 447
+- **Deletions**: 8,410
+- **Net reduction**: ~8,000 lines
+
+**Modified Files (Key)**:
+- `.opencode/agents/*.md` (7 files) - All agents refactored
+- `.opencode/skills/*/SKILL.md` (43+ files) - Skills updated
+- `.opencode/settings.json` - Settings updated
+- `scripts/file_manager.py` - Minor updates
+
+**Deleted Files** (15 - moved to claude_cli_resources/):
+- `.opencode/agents/_modules/*.md` (11 files)
+- `.opencode/agents/_domains/*.md` (3 files)
+- `.opencode/_shared/jarvis-framework-v1.md`
+
+**New Files Created** (6 script references):
+```
+claude_cli_resources/docs/scripts/
+├── context_memory/context_memory_cli.md
+├── context_memory/active_context.md
+├── rag/rag_cli.md
+├── code/lsp_navigator.md
+├── validation/policy_manager.md
+└── llm/llm_ops.md
+```
+
+### Key Accomplishments
+
+1. **Balanced Refactoring Design** - Found optimal balance between token savings and behavioral clarity
+2. **All 7 Agents Refactored** - 2242 → 532 lines (76% reduction)
+3. **4 Large Skills Refactored** - agent-invoker, jarvis-config, lsp-navigator, rag-manager
+4. **6 Script Reference Files Created** - Centralized command documentation
+5. **Agent Testing Validated** - code-reviewer agent tested successfully with full functionality
+
+### Design Principles Established
+
+1. **Clear prose instructions** - No redundant TOON for common knowledge (SOLID, DDD, INVEST)
+2. **Trust agent knowledge** - LLMs already know methodologies, just need clear instructions
+3. **Explicit criteria** - "REJECT if:" / "REQUIRED:" / "NEVER:" format
+4. **External references** - Command syntax details in separate files, loaded on-demand
+5. **Tables for structured data** - More readable than TOON for routing/language configs
+
+### Final Agent Line Counts
+
+| Agent | Original | Final | Savings |
+|-------|----------|-------|---------|
+| code-reviewer.md | 455 | 80 | 82% |
+| data-engineer.md | 230 | 68 | 70% |
+| jarvis.md | 195 | 81 | 58% |
+| product-owner.md | 314 | 87 | 72% |
+| script-developer.md | 364 | 87 | 76% |
+| script-developer-optimized.md | 330 | 56 | 83% |
+| solution-architect.md | 354 | 73 | 79% |
+| **TOTAL** | **2242** | **532** | **76%** |
+
+### What Wasn't Completed
+
+- 43 skills marked as "pending" in original scope - most already compact
+- Full validation of all 4 refactored skills (only code-reviewer agent tested)
+- Python scripts of skills not reviewed yet
+
+### Lessons Learned
+
+1. **LLMs don't need code examples of common knowledge** - Saying "follow SOLID" is enough
+2. **TOON format is overkill for simple lists** - Prose or markdown tables are clearer
+3. **Balance matters** - First refactoring (100 lines) was too thin, balanced version (80 lines) preserved clarity
+4. **Test after refactoring** - code-reviewer agent test confirmed functionality preserved
+
+### Tips for Future Developers
+
+1. When refactoring agents/skills, test them before committing
+2. Use prose for instructions, tables for structured data, code blocks for commands
+3. Don't explain what LLMs already know (SOLID, DDD, security basics)
+4. Use "REJECT if:" format for clear decision criteria
+5. External references (`@`) work well for command syntax that's rarely needed
+
+### Next Session Recommendation
+
+Review Python scripts of skills to ensure they align with the refactored SKILL.md files.

package/.opencode/devlogs/JARVIS/2026-01-15-0052-check-python-scripts-skills.md

@@ -0,0 +1,25 @@
+# Session: Check Python Scripts of Skills
+**Started**: 2026-01-15 00:52
+
+## Overview
+
+Review Python scripts associated with skills to ensure they align with refactored SKILL.md files and follow JARVIS Framework governance.
+
+## Goals
+
+1. Identify all skills that have Python scripts
+2. Verify scripts follow SOLID principles and DDD patterns
+3. Check scripts are within 400 line limit
+4. Ensure scripts align with their SKILL.md documentation
+5. Identify any inconsistencies or issues
+
+## Scope
+
+Skills with scripts to review (in `.opencode/skills/*/scripts/`):
+- To be discovered
+
+## Progress
+
+### Session Log
+
+---

package/.opencode/devlogs/JARVIS/2026-01-15-1036-fix-canonical-paths-for-core-content.md

@@ -0,0 +1,261 @@
+# Session: Fix Canonical Paths for Core Content
+
+**Started**: 2026-01-15 10:36
+**Branch**: enhancement/refactoring-squashed
+
+## Overview
+
+Critical governance issue: Multiple database files in different locations causing inconsistent access. Scripts hardcoding paths instead of using canonical `get_context_memory_db_path()`.
+
+## Goals
+
+1. Create governance domain for path management
+2. Fix scripts that hardcode paths
+3. Clean up orphan database files
+4. Establish single source of truth for all core paths
+
+## Problem Analysis
+
+### Multiple DB Files (Orphans)
+- `/context_memory.db` (root) - ORPHAN
+- `/data/context_memory.db` - ORPHAN
+- `/workspace/context_memory/context.db` - ORPHAN
+- `/data/context_memory/context.db` - CANONICAL
+
+### Scripts Hardcoding Paths
+- `context_active_context.py:39`
+- `clean_context_memory.py:145`
+- `reset_context_memory.py:17`
+- `setup_advanced.py:103`
+
+## Progress
+
+- [x] Create path governance domain
+- [x] Fix hardcoded paths in scripts
+- [x] Remove orphan DB files
+- [x] Validate all paths use registry
+
+---
+
+## Notes
+
+---
+
+### Update - 2026-01-15 11:05
+
+**Summary**: Path Governance System implemented with parallel agents
+
+**Completed (Phase 1)**:
+- ✅ PathKey enum created (`src/domain/governance/path_keys.py` - 14 keys)
+- ✅ PathRegistry class added (`src/shared/paths.py` - 163 lines)
+- ✅ `.paths.yaml` updated (37 lines, all canonical paths)
+- ✅ 5 scripts fixed to use `get_context_memory_db_path()`
+- ✅ 3 orphan DBs removed
+
+**Parallel Execution**: 4 agents completed in ~45 seconds
+
+**Git Changes**:
+- Added: `src/domain/governance/path_keys.py`, `src/domain/governance/__init__.py`
+- Modified: `src/shared/paths.py`, `.paths.yaml`
+- Modified: `scripts/context_active_context.py`, `scripts/maintenance/*.py`
+
+**Discovered (Phase 2 needed)**:
+- ~15 more files with hardcoded paths
+- Categories: RAG/Chroma (4 files), Config (3 files), Workspace (8 files)
+- Most critical: `./chroma_db_ollama` hardcoded in 4 RAG scripts
+
+**Integration Test**: PASSED
+```
+CONTEXT_MEMORY_DB: EXISTS ✅
+CONFIG_PARAMETERS: EXISTS ✅
+PROJECT_ROOT: EXISTS ✅
+PROJECT_SCRIPTS: EXISTS ✅
+```
+
+**Next**: Fix remaining 15 files with hardcoded paths
+
+---
+
+### Update - 2026-01-15 11:30
+
+**Summary**: Phase 2 Complete - All hardcoded paths fixed
+
+**Completed (Phase 2)**:
+- ✅ RAG/Chroma category (5 files):
+  - `rag_config.py`, `domain_mapper_v2.py`, `detect_duplicates.py`, `domain_mapper_config.py`, `workspace_audit.py`
+- ✅ Config category (3 files):
+  - `policy_manager.py`, `workspace_path_manager.py`, `path_manager.py`
+- ✅ Workspace category (6 files):
+  - `user_story_manager.py`, `schema_manager.py`, `install_rag_dependencies.py`, `install_rag_lite.py`, `error_router.py`, `workspace_utils.py`
+- ✅ Fixed `.paths.yaml` RAG_CHROMA_DB path (was pointing to wrong location)
+
+**Parallel Execution**: 3 agents completed simultaneously
+
+**Integration Test**: ALL 13 PathKeys PASSED
+```
+CONTEXT_MEMORY_DB: ✅
+COMMAND_HISTORY_DB: ✅
+RAG_CHROMA_DB: ✅
+DOMAIN_MAPS_DB: ✅
+CONFIG_PARAMETERS: ✅
+CONFIG_POLICIES: ✅
+CONFIG_WORKSPACE_PATHS: ✅
+CONFIG_PATHS: ✅
+PROJECT_ROOT: ✅
+PROJECT_SCRIPTS: ✅
+PROJECT_DOMAIN: ✅
+PROJECT_WORKSPACE: ✅
+PROJECT_CONFIG: ✅
+```
+
+**Total Files Fixed**: 14 files across scripts/, src/infrastructure/
+
+**Status**: COMPLETE - Path Governance System fully operational
+
+---
+
+### Update - 2026-01-15 11:45
+
+**Summary**: Data Structure Normalization + Governance Rules
+
+**Problem Identified**:
+- `chroma_db_ollama` was in project root (wrong location)
+- Having data there was result of previous mess, not correct placement
+- Governance domain should define canonical locations, not scripts
+
+**Actions Taken**:
+1. **Moved** `./chroma_db_ollama` → `data/chroma_db_ollama` (canonical location)
+2. **Updated** `.paths.yaml` to point to `data/chroma_db_ollama`
+3. **Removed** orphan `data/chroma_jarvis_v2` (unused, 167KB)
+4. **Created** `src/domain/governance/location_rules.py`:
+   - `LocationCategory` enum (DATA, CONFIG, WORKSPACE, SCRIPTS, DOMAIN)
+   - `DataLocation` enum (canonical data paths)
+   - `ConfigLocation` enum (canonical config paths)
+   - `FORBIDDEN_LOCATIONS` list (root, workspace/, scripts/, src/)
+   - `validate_location()` function for path validation
+
+**Final Canonical Structure**:
+```
+data/
+├── context_memory/     # SQLite DBs
+│   ├── context.db
+│   └── command_history.db
+├── chroma_db_ollama/   # RAG ChromaDB ← MOVED HERE
+├── domain_maps/        # Domain mapping
+└── logs/               # Log files
+```
+
+**Governance Rules**:
+- All `.db` files MUST be under `data/`
+- All `.yaml` config MUST be under `config/`
+- No data files in project root
+- Scripts store no persistent data
+
+**Exports from `src.domain.governance`**:
+- `PathKey` - enum of path keys
+- `LocationCategory`, `DataLocation`, `ConfigLocation` - location enums
+- `validate_location()` - path validation function
+- `LOCATION_RULES`, `FORBIDDEN_LOCATIONS` - rule constants
+
+---
+
+## Session End Summary
+
+**Ended**: 2026-01-15 12:00
+**Duration**: ~1h 24m
+
+### Git Summary
+
+**Commits Made**: 1
+```
+82d71dc feat(governance): implement centralized path governance system
+```
+
+**Files Changed**: 26 files (+1067, -227 lines)
+
+| Type | Files |
+|------|-------|
+| Added | `src/domain/governance/__init__.py`, `src/domain/governance/path_keys.py`, `src/domain/governance/location_rules.py`, `.opencode/sessions/*` |
+| Modified | `.paths.yaml`, `src/shared/paths.py`, `scripts/context_active_context.py`, `scripts/error_router.py`, `scripts/user_story_manager.py`, `scripts/maintenance/clean_context_memory.py`, `scripts/maintenance/reset_context_memory.py`, `scripts/managers/policy_manager.py`, `scripts/managers/workspace_path_manager.py`, `scripts/managers/path_manager.py`, `scripts/managers/schema_manager.py`, `scripts/rag/rag_config.py`, `scripts/rag/domain_mapper_v2.py`, `scripts/rag/detect_duplicates.py`, `scripts/rag/domain_mapper_config.py`, `scripts/rag/workspace_audit.py`, `scripts/setup/install_rag_dependencies.py`, `scripts/setup/install_rag_lite.py`, `src/infrastructure/rag/workspace_utils.py` |
+
+**Remaining Unstaged**: ~80 files from previous refactoring sessions (agents, skills, docs)
+
+### Tasks Summary
+
+| Status | Count |
+|--------|-------|
+| Completed | 4/4 |
+| Remaining | 0 |
+
+**Completed Tasks**:
+- [x] Create path governance domain
+- [x] Fix hardcoded paths in scripts
+- [x] Remove orphan DB files
+- [x] Validate all paths use registry
+
+### Key Accomplishments
+
+1. **Created Path Governance Domain** (`src/domain/governance/`)
+   - `PathKey` enum with 14 canonical path keys
+   - `LocationCategory`, `DataLocation`, `ConfigLocation` enums
+   - `validate_location()` function for path validation
+   - `FORBIDDEN_LOCATIONS` and `LOCATION_RULES` constants
+
+2. **Implemented PathRegistry** (`src/shared/paths.py`)
+   - `PathRegistry.get(PathKey.XXX)` for path resolution
+   - `PathRegistry.exists()` for existence check
+   - `PathRegistry.ensure_dir()` for directory creation
+   - Backward-compatible legacy functions maintained
+
+3. **Fixed 19 Scripts** with hardcoded paths
+   - Phase 1: 5 context_memory scripts
+   - Phase 2: 14 additional scripts (RAG, Config, Workspace categories)
+
+4. **Normalized Data Structure**
+   - Moved `chroma_db_ollama` from root to `data/chroma_db_ollama`
+   - Removed 4 orphan database files
+   - Established canonical `data/` structure
+
+### Problems Encountered & Solutions
+
+| Problem | Solution |
+|---------|----------|
+| Multiple DB files in different locations | Identified canonical location, removed orphans |
+| `chroma_db_ollama` in project root | Moved to `data/chroma_db_ollama`, updated `.paths.yaml` |
+| Circular dependency (PathManager vs PathRegistry) | PathManager uses base_dir calculation, not PathRegistry |
+| 15+ files with hardcoded paths discovered | Used 3 parallel agents to fix all categories |
+
+### Breaking Changes
+
+- **None** - All changes are backward compatible
+- Legacy functions (`get_context_memory_db_path()`, etc.) still work
+- Scripts that used hardcoded paths now use PathRegistry internally
+
+### Configuration Changes
+
+- `.paths.yaml` updated with all canonical paths under `data:`, `config:`, `project:` sections
+- `rag_chroma` path corrected: `${jarvis}/data/chroma_db_ollama`
+
+### Dependencies Added/Removed
+
+- **None** - Pure refactoring, no new dependencies
+
+### Lessons Learned
+
+1. **Governance domain responsibility**: Location rules belong in domain layer, not scattered in scripts
+2. **Parallel agents**: 3-4 agents can fix independent file categories simultaneously
+3. **Data normalization**: Having data in wrong location (even if functional) creates technical debt
+4. **PathKey enum**: More scalable than method-based approach for path resolution
+
+### What Wasn't Completed
+
+- Remaining ~80 unstaged files from previous sessions (agents/skills refactoring) - separate concern
+
+### Tips for Future Developers
+
+1. **Always use PathRegistry**: `from src.shared.paths import PathRegistry; from src.domain.governance.path_keys import PathKey`
+2. **Never hardcode paths**: Use `PathRegistry.get(PathKey.XXX)` for all core paths
+3. **Data files go in `data/`**: No exceptions - enforce via `validate_location()`
+4. **Config files go in `config/`**: YAML configs only
+5. **Check governance module first**: `src/domain/governance/` defines all location rules
+