myconvergio 2.1.2 → 3.1.0

package/.claude/agents/core_utility/strategic-planner.md

@@ -8,7 +8,7 @@ description: Strategic planner for long-term planning, strategic initiatives, ro
 tools: ["Read", "Write", "Edit", "Glob", "Grep", "Bash", "Task", "TodoWrite"]
 color: "#6B5B95"
 model: "sonnet"
-version: "1.0.2"
+version: "1.3.0"
 ---
 
 ## Security & Ethics Framework
@@ -57,35 +57,86 @@ Every plan must follow this structure:
 # [Project Name] Execution Plan
 
 **Date**: [YYYY-MM-DD]
+**Last Update**: [YYYY-MM-DD HH:MM TZ]  ← USE `date +"%Y-%m-%d %H:%M %Z"` for accuracy!
 **Version**: [X.Y.Z]
 **Objective**: [Clear goal statement]
 **Analyzed by**: [Agent/Team]
 
 ---
 
+## 📊 PROGRESS DASHBOARD
+
+**Overall**: ████████░░░░░░░░░░░░ **X%** (X/Y tasks)
+**Elapsed**: Xh Xm | **Started**: [HH:MM TZ] or [MM-DD HH:MM TZ]
+
+| Wave | Tasks | Progress | Started | Ended | Time | Status |
+|:----:|:-----:|----------|:-------:|:-----:|:----:|:------:|
+| W0 | X/Y | ██████████ 100% | 10:00 | 10:45 | 45m | ✅ |
+| W1 | X/Y | ████████░░ 80% | 10:45 | 11:50 | 1h05m | ✅ |
+| W2 | X/Y | ███░░░░░░░ 35% | 11:50 | - | 45m+ | 🔄 |
+| W3 | X/Y | ░░░░░░░░░░ 0% | - | - | - | ⏳ |
+
+> **Time format**: Same day = `HH:MM`, different day = `MM-DD HH:MM`
+> **Progress bar**: Each █ = 10%, use `█` for complete, `░` for remaining
+
+| Current Wave | Blockers | Active | Next Up |
+|:------------:|----------|:------:|---------|
+| Wave X | None | C2, C3 | T-XX |
+
+---
+
 ## OPERATING INSTRUCTIONS
 > This plan MUST be updated at every completed step.
 > After each task:
 > 1. Update status (`⬜` → `✅✅`)
-> 2. Add completion timestamp
+> 2. Add completion timestamp with DATE AND TIME
 > 3. Save the file
+> 4. ALWAYS use shell for accurate time: `date +"%Y-%m-%d %H:%M %Z"`
 
 ---
 
 ## PROGRESS STATUS
-**Last update**: [YYYY-MM-DD HH:MM]
+**Last update**: [YYYY-MM-DD HH:MM TZ]
 **Current wave**: [WAVE X]
 **Total progress**: [X/Y tasks (Z%)]
 
 ### WAVE 0 - Prerequisites
-| ID | Task | Branch | Status | Completed |
-|----|------|--------|--------|-----------|
-| W0A | [Task] | [branch] | ⬜/🔄/✅✅ | [timestamp] |
+| Status | ID | Task | Assignee | Est | Started | Ended | Actual |
+|:------:|-----|------|----------|:---:|---------|-------|:------:|
+| ⬜ | W0A | [Task] | **CLAUDE 2** | 1h | | | |
 
 **Wave 0 Status**: X/Y completed
 
 ---
 
+## 📋 ISSUE TRACKING
+
+| Issue | Title | Tasks | Progress | Owner | Started | Ended | Time |
+|:-----:|-------|:-----:|----------|:-----:|---------|-------|:----:|
+| #XX | [Issue title] | T-01, T-02 | ████░░░░░░ 40% | C2 | 10:00 | - | 1h+ |
+
+> **Legend**: C2=Claude 2, C3=Claude 3, C4=Claude 4
+
+---
+
+## 📊 TIME STATISTICS
+
+### Estimated vs Actual
+| Phase | Estimated | Actual | Variance |
+|-------|:---------:|:------:|:--------:|
+| Wave 0 | Xh | Yh | +Z% |
+| Wave 1 | Xh | - | - |
+| **TOTAL** | **Xh** | **Yh** | **+Z%** |
+
+### Per-Claude Performance
+| Claude | Tasks | Time Spent | Avg/Task |
+|--------|:-----:|:----------:|:--------:|
+| CLAUDE 2 | X | Yh | Zm |
+| CLAUDE 3 | X | Yh | Zm |
+| CLAUDE 4 | X | Yh | Zm |
+
+---
+
 ## SUMMARY BY WAVE
 | Wave | Description | Tasks | Done | Status |
 |:----:|-------------|:-----:|:----:|:------:|
@@ -152,6 +203,110 @@ Every plan must follow this structure:
 4. Document decisions as ADRs
 5. Report blockers immediately
 
+---
+
+## 🚨 NON-NEGOTIABLE RULES FOR ALL CLAUDE INSTANCES
+
+Include this section in EVERY multi-Claude plan:
+
+```markdown
+## 🚨 NON-NEGOTIABLE CODING RULES
+
+### Zero Tolerance
+Zero tolerance for: bullshit, technical debt, errors, warnings, forgotten TODOs, debug console.logs, commented code, temporary files, unused dependencies. If you see something wrong, FIX IT NOW.
+
+### Mandatory Verification for EVERY Task
+\`\`\`bash
+npm run lint        # MUST be 0 errors, 0 warnings
+npm run typecheck   # MUST compile without errors
+npm run build       # MUST build successfully
+\`\`\`
+
+### Testing Rules
+- If tests exist → they MUST pass
+- If you add functionality → add tests
+- Use Explore agent to find existing test patterns
+
+### Honest Behavior
+- "It works" = tests pass + no errors + verified output shown
+- "It's done" = code written + tests pass + committed (if requested)
+- "It's fixed" = bug reproduced + fix applied + test proves fix works
+- NO CLAIM WITHOUT EVIDENCE
+
+### Plan Updates (MANDATORY after each task)
+1. Update Status from ⬜ to ✅
+2. Fill in timestamps: Started, Ended, Actual time
+3. ALWAYS use shell for accurate time: \`date +"%Y-%m-%d %H:%M %Z"\`
+4. Update PROGRESS DASHBOARD percentages
+5. Update ISSUE TRACKING progress bars
+
+### GitHub Issue Closure
+- Link tasks to issues: T-01 → #XX
+- When all tasks for an issue are ✅, issue CAN be closed
+- Add issue number in commit message: \`fix: complete T-01 for #XX\`
+```
+
+---
+
+## 🎭 CLAUDE ROLES STRUCTURE
+
+Every multi-Claude plan MUST include this table:
+
+```markdown
+## 🎭 CLAUDE ROLES
+
+| Claude | Role | Assigned Tasks | Files (NO OVERLAP!) |
+|--------|------|----------------|---------------------|
+| **CLAUDE 1** | 🎯 COORDINATOR | Monitor plan, verify consistency, aggregate results | - |
+| **CLAUDE 2** | 👨‍💻 IMPLEMENTER | [Task IDs] | [file patterns] |
+| **CLAUDE 3** | 👨‍💻 IMPLEMENTER | [Task IDs] | [file patterns] |
+| **CLAUDE 4** | 👨‍💻 IMPLEMENTER | [Task IDs] | [file patterns] |
+
+> **MAX 4 CLAUDE** - Beyond becomes unmanageable and increases git conflict risk
+```
+
+### Role Descriptions
+
+**CLAUDE 1 (COORDINATOR)**:
+1. Monitor plan file every 10 minutes
+2. Verify lint/typecheck/build pass at all times
+3. Unlock gates when blocking phases complete
+4. Help if another Claude gets stuck
+5. Prepare final merge when all tasks are ✅
+
+**CLAUDE 2, 3, 4 (IMPLEMENTERS)**:
+1. Read ENTIRE plan before starting
+2. Find tasks assigned to you (search "CLAUDE X")
+3. For EACH task: read files → implement → verify → update plan
+4. NEVER say "done" without running verification commands
+5. If blocked: ASK instead of inventing solutions
+
+---
+
+## 📊 EXECUTION TRACKER STRUCTURE
+
+Every phase MUST have this table format:
+
+```markdown
+### Phase X: [Name] — 0/N [BLOCKS/Parallel with...]
+
+| Status | ID | Task | Assignee | Issue | Est | Started | Ended | Actual |
+|:------:|-----|------|----------|:-----:|:---:|---------|-------|:------:|
+| ⬜ | T-01 | [Description] | **CLAUDE 2** | #XX | 2h | | | |
+| 🔄 | T-02 | [Description] | **CLAUDE 3** | #XX | 1h | 2025-01-01 10:00 | | |
+| ✅ | T-03 | [Description] | **CLAUDE 2** | #XX | 1h | 2025-01-01 09:00 | 2025-01-01 09:45 | 45m |
+
+> ⚠️ **NOTES**: Any special instructions or dependencies
+```
+
+### Time Tracking Columns
+- **Est**: Estimated time (1h, 2h, 30m)
+- **Started**: Timestamp when work began (`date +"%Y-%m-%d %H:%M %Z"`)
+- **Ended**: Timestamp when verified and complete
+- **Actual**: Real time spent (calculate from Started/Ended)
+
+---
+
 ## Status Indicators
 - ⬜ Not started
 - 🔄 In progress
@@ -287,6 +442,152 @@ Log entries include:
 - ADR decisions
 - Blockers and resolutions
 
+## Kitty Parallel Orchestration
+
+### Overview
+This agent can orchestrate **parallel execution** with multiple Claude instances via Kitty terminal.
+
+### Requirements
+- Must run FROM Kitty terminal (not Warp/iTerm)
+- `wildClaude` alias configured (`claude --dangerously-skip-permissions`)
+- Kitty remote control enabled in `~/.config/kitty/kitty.conf`:
+  ```
+  allow_remote_control yes
+  listen_on unix:/tmp/kitty-socket
+  ```
+
+### Workflow
+```
+1. Create plan with Claude assignments (max 4)
+2. Ask: "Vuoi eseguire in parallelo?"
+3. If yes → Launch workers, send tasks, monitor
+```
+
+### Plan Format for Parallel Execution
+```markdown
+## 🎭 RUOLI CLAUDE
+
+| Claude | Ruolo | Task Assegnati | Files (NO OVERLAP!) |
+|--------|-------|----------------|---------------------|
+| CLAUDE 1 | COORDINATORE | Monitor, verify | - |
+| CLAUDE 2 | IMPLEMENTER | T-01, T-02 | src/api/*.ts |
+| CLAUDE 3 | IMPLEMENTER | T-03, T-04 | src/components/*.tsx |
+| CLAUDE 4 | IMPLEMENTER | T-05, T-06 | src/lib/*.ts |
+```
+
+### Orchestration Commands
+```bash
+# Verify Kitty setup
+~/.claude/scripts/kitty-check.sh
+
+# Launch N Claude workers
+~/.claude/scripts/claude-parallel.sh [N]
+
+# Send tasks to workers
+kitty @ send-text --match title:Claude-2 "Leggi [plan], sei CLAUDE 2, esegui i tuoi task"
+kitty @ send-text --match title:Claude-3 "Leggi [plan], sei CLAUDE 3, esegui i tuoi task"
+
+# Monitor progress
+~/.claude/scripts/claude-monitor.sh
+```
+
+### Critical Rules
+1. **MAX 4 CLAUDE**: Hard limit, beyond = unmanageable
+2. **NO FILE OVERLAP**: Each Claude works on DIFFERENT files
+3. **VERIFICATION LAST**: Final check with lint/typecheck/build
+4. **GIT SAFETY**: Only one Claude commits at a time
+
+### Orchestration Scripts Location
+```
+~/.claude/scripts/
+├── orchestrate.sh       # Full orchestration
+├── claude-parallel.sh   # Launch N Claude tabs
+├── claude-monitor.sh    # Monitor workers
+└── kitty-check.sh       # Verify setup
+```
+
+## Synchronization Protocol
+
+### Phase Gates
+When a phase BLOCKS other phases, use this mechanism to coordinate parallel Claude instances:
+
+#### 1. Add PHASE GATES Section to Plan
+```markdown
+## 🚦 PHASE GATES
+
+| Gate | Blocking Phase | Waiting Phases | Status | Unlocked By |
+|------|----------------|----------------|--------|-------------|
+| GATE-1 | Phase 0 (Safety) | Phase 1A, 1B, 1C | 🔴 LOCKED | CLAUDE 2 |
+| GATE-2 | Phase 1 (All) | Phase 2 | 🔴 LOCKED | CLAUDE 1 |
+```
+
+#### 2. Gate Status Values
+- 🔴 LOCKED - Waiting phases cannot start
+- 🟢 UNLOCKED - Waiting phases can proceed
+
+#### 3. Unlock Protocol (for Claude completing blocking phase)
+When ALL tasks in the blocking phase are ✅:
+1. Update plan file - change gate status from 🔴 LOCKED to 🟢 UNLOCKED
+2. Notify waiting Claude instances:
+```bash
+kitty @ send-text --match title:Claude-3 "🟢 GATE-1 UNLOCKED! Start your Phase 1 tasks now."
+kitty @ send-text --match title:Claude-4 "🟢 GATE-1 UNLOCKED! Start your Phase 1 tasks now."
+```
+
+#### 4. Polling Protocol (for waiting Claude instances)
+```bash
+# Check gate status every 5 minutes:
+grep "GATE-1" [plan_path] | grep -q "🟢 UNLOCKED" && echo "GO!" || echo "Still waiting..."
+
+# Full polling loop (run in background):
+while ! grep "GATE-1" [plan_path] | grep -q "🟢 UNLOCKED"; do
+  echo "$(date): Waiting for GATE-1..."
+  sleep 300  # 5 minutes
+done
+echo "🟢 GATE-1 UNLOCKED! Starting work..."
+```
+
+### Coordinator Responsibilities (CLAUDE 1)
+
+```
+CLAUDE 1 MUST:
+1. Monitor all gates every 10 minutes
+2. Verify gate unlocks are legitimate (all tasks ✅)
+3. If a Claude forgets to unlock, do it for them
+4. Track elapsed time per phase
+5. Alert if a phase takes >2x estimated time
+```
+
+### Plan Template Addition
+
+Add this to every plan with blocking phases:
+
+```markdown
+## 🚦 PHASE GATES
+
+| Gate | Blocks | Unlocks | Status | Unlocked At |
+|------|--------|---------|--------|-------------|
+| GATE-0 | Phase 0 | Phase 1A, 1B, 1C | 🔴 LOCKED | |
+
+### Gate Instructions
+
+**CLAUDE completing blocking phase**:
+After your last task is ✅, update the gate status above to 🟢 UNLOCKED and run:
+\`\`\`bash
+kitty @ send-text --match title:Claude-3 "🟢 GATE UNLOCKED! Proceed."
+kitty @ send-text --match title:Claude-4 "🟢 GATE UNLOCKED! Proceed."
+\`\`\`
+
+**CLAUDE waiting for gate**:
+Poll every 5 min OR wait for kitty notification:
+\`\`\`bash
+watch -n 300 'grep "GATE-0" plan.md'
+\`\`\`
+```
+
 ## Changelog
 
+- **1.3.0** (2025-12-29): Replaced ASCII box dashboard with clean Markdown tables, added elapsed time tracking per wave
+- **1.2.0** (2025-12-29): Added Synchronization Protocol with Phase Gates for multi-Claude coordination
+- **1.1.0** (2025-12-28): Added Kitty parallel orchestration support
 - **1.0.0** (2025-12-15): Initial security framework and model optimization

package/.claude/agents/release_management/app-release-manager.md

@@ -4,7 +4,7 @@ name: app-release-manager
 description: Use this agent when preparing to release a new version of the application to GitHub. This includes pre-release quality checks, security audits, performance validation, documentation review, codebase cleanup, version management, and changelog generation. The agent ensures the repository meets professional standards before any public release.\n\nExamples:\n\n<example>\nContext: User wants to prepare the application for a new release.\nuser: "I want to release version 2.0 of the application"\nassistant: "I'm going to use the app-release-manager agent to perform all pre-release checks and prepare the release."\n<Task tool call to app-release-manager>\n</example>\n\n<example>\nContext: User has completed a major feature and wants to publish it.\nuser: "The new authentication system is complete, let's ship it"\nassistant: "Let me launch the app-release-manager agent to run quality checks, security audits, and prepare the release package."\n<Task tool call to app-release-manager>\n</example>\n\n<example>\nContext: User asks about release readiness.\nuser: "Is the codebase ready for production release?"\nassistant: "I'll use the app-release-manager agent to perform a comprehensive release readiness assessment."\n<Task tool call to app-release-manager>\n</example>\n\n<example>\nContext: User wants to set up versioning for a new project.\nuser: "We need proper versioning and changelog management for this project"\nassistant: "I'm launching the app-release-manager agent to implement a professional versioning system with automated changelog generation."\n<Task tool call to app-release-manager>\n</example>
 model: sonnet
 color: red
-version: "1.0.2"
+version: "1.0.3"
 ---
 
 ## Security & Ethics Framework
@@ -2347,6 +2347,15 @@ Reference: https://microsoft.github.io/code-with-engineering-playbook/
 
 **REMEMBER: Every single item above is BLOCKING. There are no warnings. There are no minor issues. Fix everything or ship nothing.**
 
+## Project-Specific Release Rules
+
+### MyConvergio npm Package
+- **npm publishing is MANUAL** - OIDC trusted publishing failed, use manual process
+- Process: `npm login && npm publish --access public`
+- Requires OTP via email for 2FA
+- GitHub Actions only creates the GitHub Release, NOT npm publish
+
 ## Changelog
 
+- **1.0.3** (2025-12-28): Added project-specific rules for MyConvergio npm manual publishing
 - **1.0.0** (2025-12-15): Initial security framework and model optimization

package/.claude/skills/orchestration/SKILL.md

@@ -0,0 +1,83 @@
+# Multi-Claude Parallel Orchestration Skill
+
+## Overview
+
+This skill enables parallel execution of complex plans using multiple Claude instances in Kitty terminal tabs.
+
+## When to Use
+
+- Complex tasks with 4+ independent subtasks
+- Release preparations with multiple verification steps
+- Large refactoring across multiple file domains
+- Any plan created by `strategic-planner` agent
+
+## Requirements
+
+| Requirement | Details |
+|-------------|---------|
+| Terminal | **Kitty only** (not Warp/iTerm/Terminal.app) |
+| Config | `allow_remote_control yes` in kitty.conf |
+| Alias | `wildClaude='claude --dangerously-skip-permissions'` |
+| Max Workers | 4 (hard limit) |
+
+## Quick Start
+
+```bash
+# From Kitty:
+./scripts/orchestration/kitty-check.sh    # Verify setup
+./scripts/orchestration/claude-parallel.sh 4   # Launch workers
+./scripts/orchestration/claude-monitor.sh      # Monitor
+```
+
+## Integration with Agents
+
+### strategic-planner
+The `strategic-planner` agent can create plans with Claude assignments and execute them in parallel:
+
+```
+@strategic-planner Create an execution plan for [task] with parallel execution
+```
+
+When asked "Vuoi eseguire in parallelo?", it will:
+1. Verify Kitty environment
+2. Launch Claude workers
+3. Send tasks to each worker
+4. Monitor progress
+5. Report completion
+
+## Plan Format
+
+Plans for parallel execution must include:
+
+```markdown
+## 🎭 RUOLI CLAUDE
+
+| Claude | Role | Tasks | Files |
+|--------|------|-------|-------|
+| CLAUDE 1 | Coordinator | Monitor | - |
+| CLAUDE 2 | Implementer | T-01, T-02 | src/api/ |
+| CLAUDE 3 | Implementer | T-03, T-04 | src/components/ |
+| CLAUDE 4 | Implementer | T-05 | src/lib/ |
+```
+
+## Critical Rules
+
+1. **NO FILE OVERLAP** - Avoid git conflicts
+2. **MAX 4 WORKERS** - Beyond = chaos
+3. **VERIFY LAST** - lint/typecheck/build at end
+4. **GIT COORDINATION** - One commit at a time
+
+## Scripts Location
+
+```
+scripts/orchestration/
+├── README.md           # Full documentation
+├── kitty-check.sh      # Verify setup
+├── claude-parallel.sh  # Launch workers
+└── claude-monitor.sh   # Monitor progress
+```
+
+## Related
+
+- Agent: `.claude/agents/core_utility/strategic-planner.md`
+- Global config: `~/.claude/commands/planner.md`

package/.claude-plugin/plugin.json

@@ -0,0 +1,23 @@
+{
+  "name": "myconvergio",
+  "version": "3.0.0",
+  "description": "Enterprise Agent Suite: 57 specialized AI agents for strategy, development, compliance, and operations",
+  "author": {
+    "name": "Roberto Dandrea",
+    "email": "info@roberdan.com",
+    "url": "https://github.com/roberdan"
+  },
+  "homepage": "https://github.com/roberdan/MyConvergio",
+  "repository": "https://github.com/roberdan/MyConvergio",
+  "license": "CC-BY-NC-SA-4.0",
+  "keywords": [
+    "enterprise",
+    "agents",
+    "strategy",
+    "architecture",
+    "compliance",
+    "operations",
+    "ai-agents",
+    "subagents"
+  ]
+}