proagents 1.6.20 → 1.6.21

package/.proagents/automation/auto-decisions.md

@@ -1,535 +0,0 @@
-# Auto-Decision System
-
-Eliminate AI questions during workflow execution through pre-configured decisions.
-
----
-
-## The Problem
-
-When AI asks questions during workflow:
-```
-AI: "Should I use REST or GraphQL for this API?"
-    ↓
-[FLOW BLOCKED - Waiting for user response]
-    ↓
-Automation broken!
-```
-
-## The Solution
-
-Pre-configure all possible decisions so AI never needs to ask:
-```
-AI: Checks decision config → Uses pre-defined answer → Continues flow
-    ↓
-[NO INTERRUPTION]
-    ↓
-Automation continues!
-```
-
----
-
-## Automation Levels
-
-### Level 1: Full Automation (No Interruptions)
-```yaml
-automation:
-  level: "full"
-  user_intervention: false
-
-  behavior:
-    unknown_decisions: "use_default"
-    log_decisions: true
-    review_after_completion: true
-```
-- AI makes ALL decisions using defaults
-- Never asks questions
-- Logs all decisions for post-review
-- Best for: CI/CD pipelines, batch processing
-
-### Level 2: Semi-Automation (Critical Only)
-```yaml
-automation:
-  level: "semi"
-  user_intervention: "critical_only"
-
-  behavior:
-    unknown_decisions: "use_default"
-    pause_on:
-      - "security_decisions"
-      - "breaking_changes"
-      - "deployment"
-```
-- Auto-decides routine questions
-- Only pauses for critical decisions
-- Best for: Daily development with safety checks
-
-### Level 3: Guided (Checkpoints Only)
-```yaml
-automation:
-  level: "guided"
-  user_intervention: "checkpoints"
-
-  behavior:
-    auto_decide: true
-    pause_at_checkpoints: true
-```
-- Auto-decides within phases
-- Pauses at configured checkpoints
-- Best for: New projects, learning the workflow
-
-### Level 4: Manual (Ask Everything)
-```yaml
-automation:
-  level: "manual"
-  user_intervention: true
-```
-- Traditional mode
-- AI asks for guidance
-- Best for: Complex/sensitive projects
-
----
-
-## Pre-Configured Decision Categories
-
-### 1. Architecture Decisions
-
-```yaml
-decisions:
-  architecture:
-    # API Style
-    api_style:
-      default: "rest"
-      options: ["rest", "graphql", "grpc", "trpc"]
-
-    # State Management
-    state_management:
-      default: "zustand"
-      options: ["zustand", "redux", "context", "jotai", "mobx"]
-
-    # Styling Approach
-    styling:
-      default: "tailwind"
-      options: ["tailwind", "css-modules", "styled-components", "scss"]
-
-    # Database
-    database:
-      default: "postgresql"
-      options: ["postgresql", "mysql", "mongodb", "sqlite"]
-
-    # ORM
-    orm:
-      default: "prisma"
-      options: ["prisma", "typeorm", "drizzle", "knex"]
-
-    # Authentication
-    auth_method:
-      default: "jwt"
-      options: ["jwt", "session", "oauth", "magic-link"]
-```
-
-### 2. Code Style Decisions
-
-```yaml
-decisions:
-  code_style:
-    # Component Style
-    component_style:
-      default: "functional"
-      options: ["functional", "class"]
-
-    # File Naming
-    file_naming:
-      components: "PascalCase"
-      utilities: "camelCase"
-      constants: "UPPER_SNAKE_CASE"
-
-    # Export Style
-    export_style:
-      default: "named"
-      options: ["named", "default", "barrel"]
-
-    # Props Definition
-    props_style:
-      default: "interface"
-      options: ["interface", "type"]
-```
-
-### 3. Testing Decisions
-
-```yaml
-decisions:
-  testing:
-    # Test Framework
-    unit_framework:
-      default: "vitest"
-      options: ["vitest", "jest"]
-
-    # E2E Framework
-    e2e_framework:
-      default: "playwright"
-      options: ["playwright", "cypress"]
-
-    # Coverage Threshold
-    coverage_threshold:
-      default: 80
-      minimum: 60
-
-    # Test Location
-    test_location:
-      default: "colocated"  # Next to source files
-      options: ["colocated", "separate"]  # Or in /tests folder
-```
-
-### 4. Implementation Decisions
-
-```yaml
-decisions:
-  implementation:
-    # When multiple approaches exist
-    approach_selection:
-      default: "simplest"  # Choose simplest working solution
-      options: ["simplest", "most_performant", "most_scalable"]
-
-    # Error Handling
-    error_handling:
-      default: "try_catch"
-      options: ["try_catch", "error_boundary", "result_type"]
-
-    # Async Pattern
-    async_pattern:
-      default: "async_await"
-      options: ["async_await", "promises", "callbacks"]
-
-    # Null Handling
-    null_handling:
-      default: "optional_chaining"
-      options: ["optional_chaining", "explicit_checks", "nullish_coalescing"]
-```
-
-### 5. Deployment Decisions
-
-```yaml
-decisions:
-  deployment:
-    # Environment
-    target_environment:
-      default: "staging"
-      options: ["development", "staging", "production"]
-
-    # Strategy
-    deployment_strategy:
-      default: "rolling"
-      options: ["rolling", "blue_green", "canary"]
-
-    # Auto Deploy
-    auto_deploy:
-      staging: true
-      production: false  # Always require approval for prod
-```
-
----
-
-## Decision Rules Engine
-
-Define rules for automatic decisions:
-
-```yaml
-decision_rules:
-  # Rule: If creating API endpoint, use REST by default
-  - condition: "creating_api_endpoint"
-    decision: "api_style"
-    value: "rest"
-    unless: "graphql_already_used"
-
-  # Rule: Match existing patterns in codebase
-  - condition: "existing_pattern_found"
-    action: "use_existing_pattern"
-    priority: "high"
-
-  # Rule: For new components, use project's component style
-  - condition: "creating_component"
-    decision: "component_style"
-    value: "{{project.patterns.components.style}}"
-
-  # Rule: Use TypeScript if project uses TypeScript
-  - condition: "project_uses_typescript"
-    decision: "language"
-    value: "typescript"
-
-  # Rule: For security-related code, add extra validation
-  - condition: "security_related"
-    action: "add_validation"
-    requires_review: true
-```
-
----
-
-## Smart Defaults Based on Context
-
-### Project Type Defaults
-
-```yaml
-# Automatically set based on /init selection
-project_type_defaults:
-  web-frontend:
-    api_style: null  # No API
-    state_management: "zustand"
-    styling: "tailwind"
-    testing: "vitest"
-
-  fullstack:
-    api_style: "rest"
-    state_management: "react-query"
-    orm: "prisma"
-    auth: "next-auth"
-
-  mobile:
-    navigation: "react-navigation"
-    state_management: "zustand"
-    storage: "async-storage"
-
-  backend:
-    api_style: "rest"
-    orm: "prisma"
-    validation: "zod"
-    auth: "jwt"
-```
-
-### Codebase Analysis Defaults
-
-```yaml
-# Learn from existing codebase
-analysis_based_defaults:
-  enabled: true
-
-  learn_from:
-    - existing_patterns
-    - naming_conventions
-    - file_structure
-    - dependency_choices
-
-  apply_to:
-    - new_components
-    - new_services
-    - new_tests
-```
-
----
-
-## Question Bypass Strategies
-
-### Strategy 1: Pre-Answer Common Questions
-
-```yaml
-common_questions:
-  "Should I create a new component or modify existing?":
-    rule: "If similar component exists with >70% match, modify. Otherwise create new."
-    auto_answer: true
-
-  "Which folder should this go in?":
-    rule: "Follow existing project structure patterns."
-    auto_answer: true
-
-  "Should I add tests for this?":
-    rule: "Always yes, follow testing.test_location config."
-    auto_answer: true
-
-  "What should I name this?":
-    rule: "Follow naming conventions from decisions.code_style."
-    auto_answer: true
-
-  "Should I refactor this while I'm here?":
-    rule: "No, stick to the requested task only."
-    auto_answer: true
-```
-
-### Strategy 2: Decision Inference
-
-```yaml
-inference_rules:
-  # Infer from task description
-  - pattern: "add.*api.*endpoint"
-    infer:
-      api_style: "match_existing_or_default"
-      add_validation: true
-      add_tests: true
-
-  # Infer from file location
-  - pattern: "src/components/*"
-    infer:
-      type: "component"
-      style: "project_component_style"
-      test_location: "colocated"
-
-  # Infer from similar existing code
-  - pattern: "similar_code_exists"
-    infer:
-      follow_existing_pattern: true
-```
-
-### Strategy 3: Fallback Chain
-
-```yaml
-fallback_chain:
-  # When decision needed, try in order:
-  1. Check explicit config (decisions.yaml)
-  2. Check project type defaults
-  3. Check codebase analysis results
-  4. Check inference rules
-  5. Use global defaults
-  6. If still unknown:
-     - If automation=full: Use safest default
-     - If automation=semi: Queue for review
-     - If automation=guided: Pause and ask
-```
-
----
-
-## Decision Logging
-
-All auto-decisions are logged for review:
-
-```yaml
-decision_log:
-  enabled: true
-  location: ".proagents/decisions/"
-
-  log_format:
-    timestamp: true
-    decision_type: true
-    rule_applied: true
-    value_chosen: true
-    alternatives: true
-    confidence: true
-```
-
-### Sample Decision Log
-
-```json
-{
-  "session": "2024-01-15-feature-auth",
-  "decisions": [
-    {
-      "timestamp": "2024-01-15T10:30:00Z",
-      "question": "Which state management to use?",
-      "decision": "zustand",
-      "rule_applied": "project_type_defaults.web-frontend.state_management",
-      "confidence": "high",
-      "alternatives": ["redux", "context", "jotai"]
-    },
-    {
-      "timestamp": "2024-01-15T10:32:00Z",
-      "question": "Where to place AuthService?",
-      "decision": "src/services/authService.ts",
-      "rule_applied": "codebase_analysis.existing_pattern",
-      "confidence": "high",
-      "alternatives": []
-    },
-    {
-      "timestamp": "2024-01-15T10:35:00Z",
-      "question": "Add error boundary?",
-      "decision": "yes",
-      "rule_applied": "common_questions.error_handling",
-      "confidence": "medium",
-      "alternatives": ["no", "custom_error_handler"]
-    }
-  ],
-  "summary": {
-    "total_decisions": 15,
-    "auto_resolved": 15,
-    "user_intervention": 0,
-    "confidence_breakdown": {
-      "high": 12,
-      "medium": 3,
-      "low": 0
-    }
-  }
-}
-```
-
----
-
-## Post-Execution Review
-
-After automated workflow completes:
-
-```
-┌─────────────────────────────────────────────────────────┐
-│ Workflow Complete - Decision Review                     │
-├─────────────────────────────────────────────────────────┤
-│ Feature: user-authentication                            │
-│ Duration: 45 minutes                                    │
-│ Automation Level: Full                                  │
-├─────────────────────────────────────────────────────────┤
-│ Decisions Made: 23                                      │
-│ ├── High Confidence: 18                                │
-│ ├── Medium Confidence: 4                               │
-│ └── Low Confidence: 1 ⚠️                               │
-├─────────────────────────────────────────────────────────┤
-│ Low Confidence Decision:                                │
-│ • Used REST for user-preferences endpoint              │
-│   Reason: No clear pattern, used default               │
-│   Action: Review if GraphQL would be better            │
-├─────────────────────────────────────────────────────────┤
-│ [View Full Log] [Approve All] [Review Flagged]         │
-└─────────────────────────────────────────────────────────┘
-```
-
----
-
-## Configuration Example
-
-### Full Automation Config
-
-```yaml
-# proagents.config.yaml
-
-automation:
-  level: "full"
-  user_intervention: false
-
-  # Pre-configured decisions
-  decisions:
-    architecture:
-      api_style: "rest"
-      state_management: "zustand"
-      styling: "tailwind"
-      database: "postgresql"
-      orm: "prisma"
-
-    code_style:
-      component_style: "functional"
-      export_style: "named"
-
-    testing:
-      framework: "vitest"
-      coverage: 80
-
-    implementation:
-      approach: "simplest"
-      error_handling: "try_catch"
-
-  # Behavior
-  behavior:
-    unknown_decisions: "use_safest_default"
-    log_all_decisions: true
-    review_after_completion: true
-
-  # Safety
-  safety:
-    pause_on_security: false  # Set true for extra safety
-    pause_on_breaking: false
-    pause_on_deployment: false
-```
-
----
-
-## Commands
-
-| Command | Description |
-|---------|-------------|
-| `pa:auto-config` | Configure automation settings |
-| `pa:auto-level [level]` | Set automation level |
-| `pa:decisions` | View/edit pre-configured decisions |
-| `pa:decision-log` | View decision log |
-| `pa:review-decisions` | Review flagged decisions |