moai-adk 0.15.1__py3-none-any.whl → 0.32.8__py3-none-any.whl

moai_adk/templates/.claude/skills/moai-alfred-workflow/SKILL.md

@@ -1,288 +0,0 @@
----
-name: moai-alfred-workflow
-version: 1.0.0
-created: 2025-11-02
-updated: 2025-11-02
-status: active
-description: Guide 4-step workflow execution with task tracking and quality gates
-keywords: ['workflow', 'execution', 'planning', 'task-tracking', 'quality']
-allowed-tools:
-  - Read
----
-
-# Alfred 4-Step Workflow Guide
-
-## Skill Metadata
-
-| Field | Value |
-| ----- | ----- |
-| **Skill Name** | moai-alfred-workflow |
-| **Version** | 1.0.0 (2025-11-02) |
-| **Status** | Active |
-| **Tier** | Alfred |
-| **Purpose** | Guide systematic 4-step workflow execution |
-
----
-
-## What It Does
-
-Alfred uses a consistent 4-step workflow for all user requests to ensure clarity, planning, transparency, and traceability.
-
-**Key capabilities**:
-- ✅ Intent clarification with questions
-- ✅ Task planning and decomposition
-- ✅ Transparent progress tracking with TodoWrite
-- ✅ Automated reporting and commits
-- ✅ Quality gate validation
-
----
-
-## When to Use
-
-**Automatic triggers**:
-- User request received → analyze intent
-- Multiple interpretation possible → use AskUserQuestion
-- Task complexity > 1 step → invoke Plan Agent
-- Executing tasks → activate TodoWrite tracking
-- Task completion → generate report
-
-**Manual reference**:
-- Understanding workflow execution
-- Planning multi-step features
-- Learning best practices for task tracking
-
----
-
-## The 4-Step Workflow
-
-### Step 1: Intent Understanding
-
-**Goal**: Clarify user intent before any action
-
-**Actions**:
-- Evaluate request clarity
-  - HIGH clarity → Skip to Step 2 directly
-  - MEDIUM/LOW clarity → Invoke AskUserQuestion
-- Present 3-5 clear options (not open-ended)
-- Gather user responses before proceeding
-
-**When to Ask Questions**:
-- Multiple tech stack choices available
-- Architecture decisions needed
-- Business/UX decisions unclear
-- Ambiguous requirements
-- Existing component impacts unknown
-
-**Example**:
-```
-User says: "Add authentication"
-          ↓
-Clarity = MEDIUM (multiple approaches possible)
-          ↓
-Ask: "Which authentication method?"
-- Option 1: JWT tokens
-- Option 2: OAuth 2.0
-- Option 3: Session-based
-```
-
----
-
-### Step 2: Plan Creation
-
-**Goal**: Analyze tasks and identify execution strategy
-
-**Actions**:
-- Invoke Plan Agent (built-in Claude agent) to:
-  - Decompose tasks into structured steps
-  - Identify dependencies between tasks
-  - Determine single vs parallel execution
-  - Estimate file changes and scope
-- Output structured task breakdown
-
-**Plan Output Format**:
-```
-Task Breakdown:
-
-Phase 1: Preparation (30 mins)
-├─ Task 1: Set up environment
-├─ Task 2: Install dependencies
-└─ Task 3: Create test fixtures
-
-Phase 2: Implementation (2 hours)
-├─ Task 4: Core feature (parallel ready)
-├─ Task 5: API endpoints (parallel ready)
-└─ Task 6: Tests (depends on 4, 5)
-
-Phase 3: Verification (30 mins)
-├─ Task 7: Integration testing
-├─ Task 8: Documentation
-└─ Task 9: Code review
-```
-
----
-
-### Step 3: Task Execution
-
-**Goal**: Execute tasks with transparent progress tracking
-
-**Actions**:
-1. Initialize TodoWrite with all tasks (status: pending)
-2. For each task:
-   - Update TodoWrite: pending → **in_progress**
-   - Execute task (call appropriate sub-agent or action)
-   - Update TodoWrite: in_progress → **completed**
-3. Handle blockers: Keep in_progress, create new blocking task
-
-**TodoWrite Rules**:
-- Each task must have:
-  - `content`: Imperative form ("Run tests", "Fix bug")
-  - `activeForm`: Present continuous ("Running tests", "Fixing bug")
-  - `status`: One of pending/in_progress/completed
-- **EXACTLY ONE** task in_progress at a time (unless Plan Agent approved parallel)
-- Mark completed ONLY when fully done (tests pass, no errors, implementation complete)
-
-**Example TodoWrite Progression**:
-
-Initial state (all pending):
-```
-1. [pending] Set up environment
-2. [pending] Install dependencies
-3. [pending] Implement core feature
-4. [pending] Write tests
-5. [pending] Documentation
-```
-
-After starting Task 1:
-```
-1. [in_progress] Set up environment     ← ONE task in progress
-2. [pending] Install dependencies
-3. [pending] Implement core feature
-4. [pending] Write tests
-5. [pending] Documentation
-```
-
-After completing Task 1 and starting Task 2:
-```
-1. [completed] Set up environment      ✅
-2. [in_progress] Install dependencies  ← NOW in progress
-3. [pending] Implement core feature
-4. [pending] Write tests
-5. [pending] Documentation
-```
-
-**Handling Blockers**:
-
-If blocked during execution:
-```
-Example: Task 4 blocked by missing library
-
-Action:
-├─ Keep Task 4 as in_progress
-├─ Create new task: "Install library X"
-├─ Add to todo list
-└─ Start new task first
-```
-
----
-
-### Step 4: Report & Commit
-
-**Goal**: Document work and create git history
-
-**Actions**:
-- **Report Generation**: ONLY if user explicitly requested
-  - ❌ Don't auto-generate in project root
-  - ✅ OK to generate in `.moai/docs/`, `.moai/reports/`, `.moai/analysis/`
-- **Git Commit**: ALWAYS create commits (mandatory)
-  - Call git-manager for all Git operations
-  - TDD commits: RED → GREEN → REFACTOR
-  - Include Alfred co-authorship
-
-**Report Conditions**:
-
-```
-User says: "Show me a report"
-         ↓
-Generate report → .moai/reports/task-completion-001.md
-
-User says: "I'm done with feature X"
-         ↓
-NO auto-report → just create commit
-(Only create report if explicitly requested)
-```
-
-**Commit Message Format**:
-```
-feat: Add authentication support
-
-- JWT token validation implemented
-- Session management added
-- Rate limiting configured
-
-🎩 Generated with Claude Code
-
-Co-Authored-By: 🎩 Alfred@[MoAI](https://adk.mo.ai.kr)
-```
-
----
-
-## Workflow Validation Checklist
-
-Before considering workflow complete:
-- ✅ All steps followed in order (Intent → Plan → Execute → Commit)
-- ✅ No assumptions made (AskUserQuestion used when unclear)
-- ✅ TodoWrite tracks all tasks
-- ✅ Reports only generated on explicit request
-- ✅ Commits created for all completed work
-- ✅ Quality gates passed (tests, linting, type checking)
-
----
-
-## Decision Trees
-
-### When to Use AskUserQuestion
-
-```
-Request clarity unclear?
-├─ YES → Use AskUserQuestion
-│   ├─ Present 3-5 clear options
-│   ├─ Use structured format
-│   └─ Wait for user response
-└─ NO → Proceed to planning
-```
-
-### When to Mark Task Completed
-
-```
-Task marked in_progress?
-├─ Code implemented → tests pass?
-├─ Tests pass → type checking pass?
-├─ Type checking pass → linting pass?
-└─ All pass → Mark COMPLETED ✅
-   └─ NOT complete → Keep in_progress ⏳
-```
-
-### When to Create Blocking Task
-
-```
-Task execution blocked?
-├─ External dependency missing?
-├─ Pre-requisite not done?
-├─ Unknown issue?
-└─ YES → Create blocking task
-   └─ Add to todo list
-   └─ Execute blocking task first
-   └─ Return to original task
-```
-
----
-
-## Key Principles
-
-1. **Clarity First**: Never assume intent
-2. **Systematic**: Follow 4 steps in order
-3. **Transparent**: Track all progress visually
-4. **Traceable**: Document every decision
-5. **Quality**: Validate before completion
-
----

moai_adk/templates/.claude/skills/moai-cc-agents/SKILL.md

@@ -1,269 +0,0 @@
----
-name: moai-cc-agents
-description: "Creating and Managing Sub-agents in Claude Code. Design agent personas, define proactive triggers, set tool permissions, structure agent files. Use when building specialized agents for code review, debugging, architecture, or domain-specific tasks."
-allowed-tools: "Read, Write, Edit, Glob, Bash"
----
-
-## Skill Metadata
-
-| Field | Value |
-| ----- | ----- |
-| Version | 1.0.0 |
-| Tier | Ops |
-| Auto-load | When creating or managing sub-agents |
-
-## What It Does
-
-Sub-agent 생성 및 관리를 위한 전체 가이드를 제공합니다. Agent persona 설계, proactive trigger 정의, tool 권한 설정, agent 파일 구조화 방법을 다룹니다.
-
-## When to Use
-
-- Sub-agent를 새로 생성할 때
-- 기존 agent의 권한이나 trigger 조건을 수정할 때
-- Agent 간 협업 패턴을 설계할 때
-- Tool access 최소화 원칙을 적용할 때
-
-
-# Creating and Managing Sub-agents
-
-Sub-agents are specialized Claude instances with independent context, custom prompts, and restricted tool access. They handle deep analysis, parallel work, and autonomous tasks.
-
-## Agent File Structure
-
-**Location**: `.claude/agents/`
-
-```yaml
----
-name: agent-name
-description: Use PROACTIVELY for [specific trigger conditions]
-tools: Read, Write, Edit, Glob, Grep, Bash(git:*)
-model: sonnet
----
-
-# Agent Name — Specialist Role
-
-Brief description of agent expertise.
-
-## Core Mission
-
-- Primary responsibility
-- Scope boundaries
-- Success criteria
-
-## Proactive Triggers
-
-- When to activate automatically
-- Specific conditions for invocation
-- Integration with workflow
-
-## Workflow Steps
-
-1. Input validation
-2. Task execution
-3. Output verification
-4. Handoff to next agent (if applicable)
-
-## Constraints
-
-- What NOT to do
-- Delegation rules
-- Quality gates
-```
-
-## Agent Persona Design Pattern
-
-```markdown
----
-name: code-reviewer
-description: Use PROACTIVELY for code review requests, PR analysis, or quality checks
-tools: Read, Glob, Grep, Bash(git:*)
-model: sonnet
----
-
-# Code Reviewer — Quality Expert
-
-Specialized in identifying code quality issues, security risks, and architecture concerns.
-
-## Core Mission
-
-- Review code for SOLID principles
-- Identify code smells and anti-patterns
-- Verify security best practices
-- Suggest improvements with rationale
-
-## Proactive Triggers
-
-- When user mentions "review", "quality", "audit"
-- After significant code changes
-- Before PR merge
-
-## Workflow Steps
-
-1. **Analyze**: Read changed files, understand context
-2. **Evaluate**: Check against TRUST 5 principles
-3. **Report**: List findings with severity & fix suggestions
-4. **Recommend**: Suggest refactoring or architectural improvements
-
-## Constraints
-
-- No direct edits (only suggestions)
-- Focus on maintainability, not style
-- Respect existing architecture decisions
-```
-
-## High-Freedom: Agent Principles
-
-### Autonomy & Expertise
-- Each agent owns 1 specialization (not 3+)
-- Define clear "when to activate" triggers
-- Agents should make decisions independently
-
-### Tool Access Minimization
-- Grant only necessary tools per role
-- Restrict Bash to specific commands: `Bash(git:*)`, `Bash(python:*)`
-- Never grant `Bash(*)`; always specify pattern
-
-### Handoff & Collaboration
-- Agent A completes → hands off to Agent B
-- Use clear "next agent" instructions
-- No circular dependencies
-
-## Medium-Freedom: Common Agent Patterns
-
-### Pattern 1: Debugger Agent
-```yaml
-name: debugger
-description: Use PROACTIVELY for error diagnosis, test failures, exception analysis
-tools: Read, Grep, Glob, Bash(pytest:*), Bash(git:*)
-model: sonnet
-```
-
-**Mission**: Diagnose errors, provide fix-forward guidance
-**Triggers**: `error`, `failed`, `exception`, `debug`
-**Output**: Root cause + suggested fixes
-
-### Pattern 2: Architect Agent
-```yaml
-name: architect
-description: Use PROACTIVELY for system design, refactoring, scalability concerns
-tools: Read, Glob, Grep, Bash(ls:*), Bash(find:*)
-model: sonnet
-```
-
-**Mission**: Design systems, propose architectures
-**Triggers**: `architecture`, `refactor`, `design`, `scalability`
-**Output**: Design document + implementation roadmap
-
-### Pattern 3: Security Agent
-```yaml
-name: security-auditor
-description: Use PROACTIVELY for vulnerability assessment, OWASP checks, secrets detection
-tools: Read, Glob, Grep
-model: sonnet
-```
-
-**Mission**: Find security issues, verify OWASP compliance
-**Triggers**: `security`, `audit`, `vulnerability`, `secrets`
-**Output**: Risk assessment + remediation steps
-
-## Low-Freedom: Tool Permission Patterns
-
-### Principle of Least Privilege
-
-```yaml
-# ❌ Too permissive
-tools: Read, Write, Edit, Bash(*)
-
-# ✅ Appropriate
-tools: Read, Glob, Grep  # Read-only analysis
-
-# ✅ Appropriate
-tools: Read, Edit, Bash(black:*), Bash(pytest:*)  # Formatter + tests only
-```
-
-### Bash Command Restrictions
-
-```yaml
-# Allowed patterns:
-Bash(git:*)           # All git commands
-Bash(npm run:*)       # Only npm run scripts
-Bash(python:*)        # Python interpreter
-Bash(pytest:*)        # Pytest runner
-
-# Denied patterns:
-Bash(rm:*)            # Dangerous deletion
-Bash(sudo:*)          # Privilege escalation
-Bash(curl:*)          # Arbitrary downloads
-```
-
-## Agent Execution Modes
-
-### Mode 1: Inline (main context)
-```bash
-/role security
-"Check this project for vulnerabilities"
-# Runs in main session context
-```
-
-### Mode 2: Sub-agent (isolated context)
-```bash
-/role security --agent
-"Perform comprehensive security audit"
-# Runs in independent context, can work in parallel
-```
-
-### Mode 3: Parallel Multi-role
-```bash
-/multi-role security,performance,qa --agent
-"Analyze security, performance, and quality"
-# All roles run in parallel, results merged
-```
-
-## Agent Registration & Discovery
-
-```bash
-# List available agents
-/agents
-
-# Create new agent interactively
-/agents create
-
-# View specific agent
-/agents view security-auditor
-
-# Edit agent
-/agents edit security-auditor
-
-# Delete agent
-/agents delete security-auditor
-```
-
-## Agent Validation Checklist
-
-- [ ] `name` is kebab-case (e.g., `code-reviewer`)
-- [ ] `description` includes "Use PROACTIVELY for"
-- [ ] `tools` list only necessary tools
-- [ ] No `Bash(*)` wildcard; specific patterns only
-- [ ] `model` is `haiku` or `sonnet`
-- [ ] `Proactive Triggers` section clearly defined
-- [ ] No overlapping responsibilities with other agents
-- [ ] YAML frontmatter is valid
-
-## Best Practices
-
-✅ **DO**:
-- Design agents around specific expertise
-- Use `--agent` flag for large analyses
-- Combine multiple agents for complex tasks
-- Name agents descriptively
-
-❌ **DON'T**:
-- Create overpowered agents with all tools
-- Allow direct file modifications without approval
-- Overlap agent responsibilities
-- Use `Bash(*)` without specific patterns
-
----
-
-**Reference**: Claude Code Sub-agents documentation
-**Version**: 1.0.0

moai_adk/templates/.claude/skills/moai-cc-agents/templates/agent-template.md

@@ -1,32 +0,0 @@
----
-name: agent-name
-description: Use PROACTIVELY for [specific trigger conditions]
-tools: Read, Glob, Grep, Bash(git:*)
-model: sonnet
----
-
-# Agent Name — Specialist Role
-
-Brief description of agent expertise.
-
-## Core Mission
-
-- Primary responsibility
-- Scope boundaries
-- Success criteria
-
-## Proactive Triggers
-
-- When to activate automatically
-- Specific conditions for invocation
-
-## Workflow Steps
-
-1. Input validation
-2. Task execution
-3. Output verification
-
-## Constraints
-
-- What NOT to do
-- Delegation rules