pdca-tracker

name	pdca-tracker
description	PDCA cycle tracking skill for plan-do-check-act improvement management.
allowed-tools	Bash(*) Read Write Edit Glob Grep WebFetch
metadata	{"author":"babysitter-sdk","version":"1.0.0","category":"continuous-improvement","backlog-id":"SK-IE-041"}
graph	{"domains":["domain:industrial-engineering"],"skillAreas":["skill-area:statistical-analysis","skill-area:organizational-design","skill-area:data-analysis"],"roles":["role:operations-analyst","role:research-engineer"]}

pdca-tracker

You are pdca-tracker - a specialized skill for tracking PDCA (Plan-Do-Check-Act) cycles and improvement management.

Overview

This skill enables AI-powered PDCA tracking including:

• PDCA cycle setup and management
• Hypothesis development
• Experiment planning
• Results verification
• Standard work updates
• Cycle iteration tracking
• Learning documentation
• Multi-project portfolio view

Capabilities

1. PDCA Cycle Setup

from dataclasses import dataclass
from typing import List, Dict, Optional
from datetime import datetime, timedelta
from enum import Enum
import uuid

class PDCAPhase(Enum):
    PLAN = "plan"
    DO = "do"
    CHECK = "check"
    ACT = "act"

@dataclass
class PDCACycle:
    id: str
    title: str
    owner: str
    start_date: datetime
    current_phase: PDCAPhase
    iteration: int = 1

def create_pdca_cycle(title: str, owner: str, hypothesis: str,
                     success_criteria: Dict):
    """
    Create new PDCA cycle

    hypothesis: What we believe will happen
    success_criteria: Measurable criteria for success
    """
    cycle_id = str(uuid.uuid4())[:8]

    cycle = {
        "id": cycle_id,
        "title": title,
        "owner": owner,
        "created_date": datetime.now().strftime("%Y-%m-%d"),
        "iteration": 1,
        "current_phase": "PLAN",
        "phases": {
            "PLAN": {
                "status": "in_progress",
                "hypothesis": hypothesis,
                "success_criteria": success_criteria,
                "planned_actions": [],
                "resources_needed": [],
                "timeline": None,
                "completed_date": None
            },
            "DO": {
                "status": "not_started",
                "actions_taken": [],
                "observations": [],
                "data_collected": [],
                "issues_encountered": [],
                "completed_date": None
            },
            "CHECK": {
                "status": "not_started",
                "results": {},
                "hypothesis_validated": None,
                "learnings": [],
                "completed_date": None
            },
            "ACT": {
                "status": "not_started",
                "decision": None,  # standardize, adjust, abandon
                "standard_work_updates": [],
                "next_cycle_needed": None,
                "completed_date": None
            }
        },
        "history": []
    }

    return cycle

2. Plan Phase Management

def develop_plan(cycle: Dict, plan_details: Dict):
    """
    Develop the Plan phase

    plan_details: {
        'actions': [{'description': str, 'owner': str, 'due_date': str}],
        'timeline': {'start': str, 'end': str},
        'resources': [str],
        'risks': [str]
    }
    """
    cycle['phases']['PLAN']['planned_actions'] = plan_details.get('actions', [])
    cycle['phases']['PLAN']['timeline'] = plan_details.get('timeline')
    cycle['phases']['PLAN']['resources_needed'] = plan_details.get('resources', [])
    cycle['phases']['PLAN']['risks'] = plan_details.get('risks', [])

    # Validate plan completeness
    validation = validate_plan(cycle['phases']['PLAN'])

    if validation['is_complete']:
        cycle['phases']['PLAN']['status'] = 'complete'
        cycle['phases']['PLAN']['completed_date'] = datetime.now().strftime("%Y-%m-%d")
        cycle['current_phase'] = 'DO'
        cycle['phases']['DO']['status'] = 'in_progress'

        # Log transition
        cycle['history'].append({
            'timestamp': datetime.now().isoformat(),
            : ,
            : ,
            : 
        })

     {
        : cycle,
        : validation
    }

 ():
    
    issues = []

      plan.get():
        issues.append()
      plan.get():
        issues.append()
      plan.get():
        issues.append()
      plan.get():
        issues.append()

     {
        : (issues) == ,
        : issues
    }

3. Do Phase Tracking

def track_do_phase(cycle: Dict, execution_data: Dict):
    """
    Track execution in Do phase

    execution_data: {
        'action_id': str,
        'status': str,
        'observations': [str],
        'data_points': [{'metric': str, 'value': float, 'timestamp': str}],
        'issues': [str]
    }
    """
    do_phase = cycle['phases']['DO']

    # Update action status
    for action in cycle['phases']['PLAN']['planned_actions']:
        if action.get('id') == execution_data.get('action_id'):
            action['status'] = execution_data['status']
            action['actual_completion'] = datetime.now().strftime("%Y-%m-%d")

    # Record observations
    if execution_data.get('observations'):
        do_phase['observations'].extend(execution_data['observations'])

    # Collect data
    if execution_data.get('data_points'):
        do_phase['data_collected'].extend(execution_data['data_points'])

    # Record issues
    if execution_data.get('issues'):
        do_phase['issues_encountered'].extend(execution_data['issues'])

    # Check if Do phase is complete
    planned_actions = cycle['phases']['PLAN']['planned_actions']
    completed = sum(1  a  planned_actions  a.get() == )

     completed == (planned_actions):
        do_phase[] = 
        do_phase[] = datetime.now().strftime()
        cycle[] = 
        cycle[][][] = 

        cycle[].append({
            : datetime.now().isoformat(),
            : ,
            : ,
            : 
        })

     {
        : cycle,
        : {
            : completed,
            : (planned_actions),
            : (do_phase[]),
            : (do_phase[])
        }
    }

4. Check Phase Analysis

import numpy as np

def analyze_check_phase(cycle: Dict):
    """
    Analyze results in Check phase
    """
    check_phase = cycle['phases']['CHECK']
    plan_phase = cycle['phases']['PLAN']
    do_phase = cycle['phases']['DO']

    results = {}

    # Compare results to success criteria
    success_criteria = plan_phase['success_criteria']
    data_collected = do_phase['data_collected']

    criteria_results = []
    for criterion, target in success_criteria.items():
        # Get data for this metric
        metric_data = [d['value'] for d in data_collected if d['metric'] == criterion]

        if metric_data:
            actual = np.mean(metric_data)
            met = (actual >= target if isinstance(target, (int, float))
                   else str(actual) == str(target))

            criteria_results.append({
                'criterion': criterion,
                'target': target,
                'actual': round(actual, 2) if isinstance(actual, float) else actual,
                'met': met
            })

    # Validate hypothesis
    criteria_met = sum(  c  criteria_results  c[])
    total_criteria = (criteria_results)

    hypothesis_validated = criteria_met == total_criteria  total_criteria >   

    check_phase[] = {
        : criteria_results,
        : criteria_met,
        : total_criteria
    }
    check_phase[] = hypothesis_validated

    
    learnings = generate_learnings(criteria_results, do_phase[],
                                   do_phase[])
    check_phase[] = learnings

     {
        : cycle,
        : {
            : hypothesis_validated,
            : (criteria_met / total_criteria * , )  total_criteria >   ,
            : criteria_results,
            : learnings
        }
    }

 ():
    
    learnings = []

    
     cr  criteria_results:
         cr[]:
            learnings.append()
        :
            learnings.append()

    
     issues:
        learnings.append()

     learnings

5. Act Phase Decision

def complete_act_phase(cycle: Dict, decision: str, next_steps: Dict):
    """
    Complete Act phase with decision

    decision: 'standardize', 'adjust', 'abandon'
    next_steps: {
        'standard_work_updates': [str],
        'next_cycle_hypothesis': str,  # if adjust
        'reason_for_abandonment': str   # if abandon
    }
    """
    act_phase = cycle['phases']['ACT']

    act_phase['decision'] = decision

    if decision == 'standardize':
        act_phase['standard_work_updates'] = next_steps.get('standard_work_updates', [])
        act_phase['next_cycle_needed'] = False

        # Mark cycle complete
        act_phase['status'] = 'complete'
        act_phase['completed_date'] = datetime.now().strftime("%Y-%m-%d")
        cycle['status'] = 'completed'

        cycle['history'].append({
            'timestamp': datetime.now().isoformat(),
            'event': 'cycle_complete',
            'decision': 'standardize',
            'outcome': 'success'
        })

    elif decision == 'adjust':
        act_phase['next_cycle_needed'] = True
        act_phase['next_hypothesis'] = next_steps.get('next_cycle_hypothesis')
        act_phase['adjustments'] = next_steps.get('adjustments', [])

        # Prepare next iteration
        cycle['iteration'] += 
        new_cycle = prepare_next_iteration(cycle)

        cycle[].append({
            : datetime.now().isoformat(),
            : ,
            : cycle[],
            : act_phase[]
        })

         {: cycle, : new_cycle}

     decision == :
        act_phase[] = next_steps.get()
        act_phase[] = 

        act_phase[] = 
        cycle[] = 

        cycle[].append({
            : datetime.now().isoformat(),
            : ,
            : act_phase[]
        })

     {: cycle}

 ():
    
     {
        : current_cycle[],
        : current_cycle[][].get(),
        : current_cycle[][][],
        : current_cycle[][].get(, [])
    }

6. PDCA Portfolio View

def get_portfolio_status(cycles: List[Dict]):
    """
    Get portfolio view of all PDCA cycles
    """
    summary = {
        'total_cycles': len(cycles),
        'by_phase': {phase.value: 0 for phase in PDCAPhase},
        'by_status': {'active': 0, 'completed': 0, 'abandoned': 0},
        'iterations': [],
        'cycle_details': []
    }

    for cycle in cycles:
        # Count by phase
        current_phase = cycle.get('current_phase', 'PLAN')
        summary['by_phase'][current_phase.lower()] += 1

        # Count by status
        status = cycle.get('status', 'active')
        summary['by_status'][status] += 1

        # Track iterations
        summary['iterations'].append(cycle.get('iteration', 1))

        # Cycle summary
        summary['cycle_details'].append({
            'id': cycle['id'],
            'title': cycle['title'],
            'owner': cycle['owner'],
            'phase': current_phase,
            'iteration': cycle.get(, ),
            : status,
            : cycle[][].get()
        })

    
    summary[] = (np.mean(summary[]), )  summary[]  
    summary[] = (
        (  c  cycles  c[][].get() == ) /
        (cycles) * , 
    )  cycles  

     summary

7. Learning Documentation

def document_learnings(cycle: Dict):
    """
    Create comprehensive learning document from PDCA cycle
    """
    learning_doc = {
        'cycle_id': cycle['id'],
        'title': cycle['title'],
        'date_completed': cycle['phases']['ACT'].get('completed_date'),
        'iterations': cycle.get('iteration', 1),
        'sections': {
            'hypothesis': cycle['phases']['PLAN']['hypothesis'],
            'what_we_tried': [],
            'what_happened': [],
            'what_we_learned': [],
            'what_changed': [],
            'recommendations': []
        }
    }

    # What we tried
    for action in cycle['phases']['PLAN']['planned_actions']:
        learning_doc['sections']['what_we_tried'].append(action['description'])

    # What happened
    for obs in cycle['phases']['DO']['observations']:
        learning_doc['sections']['what_happened'].append(obs)

    for issue in cycle['phases']['DO'][]:
        learning_doc[][].append()

    
    learning_doc[][] = cycle[][][]

    
     cycle[][][] == :
        learning_doc[][] = cycle[][][]
     cycle[][][] == :
        learning_doc[][] = [
            
        ]

    
    learning_doc[][] = generate_recommendations(cycle)

     learning_doc

 ():
    
    recommendations = []

     cycle[][].get():
        recommendations.append()
        recommendations.append()
    :
        recommendations.append()
         cycle.get(, ) >= :
            recommendations.append()

     recommendations

Process Integration

This skill integrates with the following processes:

• continuous-improvement-program.js
• a3-problem-solving-project.js
• kaizen-event-execution.js

Output Format

{
  "pdca_cycle": {
    "id": "abc12345",
    "title": "Reduce Setup Time",
    "current_phase": "CHECK",
    "iteration": 2
  },
  "plan": {
    "hypothesis": "Standardized tool staging will reduce setup 25%",
    "success_criteria": {"setup_time_minutes": 15}
  },
  "check": {
    "hypothesis_validated": false,
    "actual": 18,
    "gap": 3
  },
  "act": {
    "decision" 
     
  

Best Practices

1. Start with hypothesis - Clear, testable statement
2. Define success criteria - Measurable before starting
3. Small experiments - Test quickly, learn fast
4. Document everything - Learning is the product
5. Iterate deliberately - Each cycle builds on previous
6. Share learnings - Others benefit from your experiments

Constraints

• Requires discipline to follow process
• Not for emergencies requiring immediate action
• Data collection takes time
• Multiple iterations may be needed