Continuous Improvement & Kaizen

TL;DR: Implements systematic continuous improvement through retrospective analysis, PDCA cycles, kaizen events, and process optimization. Identifies improvement opportunities from project data, quality audits, team feedback, and lessons learned. Prioritizes improvements by effort-impact ratio and tracks implementation.

Principio Rector

La mejora continua no es un evento — es una disciplina. Cada sprint, cada fase, cada entregable es una oportunidad para mejorar el proceso que lo produjo. La clave no es la cantidad de mejoras identificadas sino la cantidad implementadas y verificadas. Una mejora no verificada es una intención, no una mejora.

Assumptions & Limits

• Assumes retrospective or feedback data exists to identify improvement opportunities [SUPUESTO]
• Assumes team has capacity to implement improvements alongside delivery work [PLAN]
• Breaks when improvement backlog grows indefinitely without implementation — signals deeper issue
• Root cause analysis requires honest team participation; blame culture produces sanitized inputs
• Does not implement organizational process changes — those require PMO-level authority [STAKEHOLDER]
• Improvement verification requires before/after metrics; without metrics, improvement is assumed

Usage

# Identify improvement opportunities from retrospective data
/pm:continuous-improvement $PROJECT --type=identify --source="retrospectives"

# Run PDCA cycle for a specific improvement
/pm:continuous-improvement $PROJECT --type=pdca --improvement="reduce-rework"

# Track improvement implementation across the project
/pm:continuous-improvement $PROJECT --type=track --status="all"

Parameters:

Parameter	Required	Description
$PROJECT	Yes	Project identifier
--type	Yes	identify, pdca, track, verify, share
--source	No	Data source (retrospectives, audits, metrics, feedback)
--improvement	No	Specific improvement to manage
--status	No	Filter by status (open, in-progress, verified, standardized)

Service Type Routing

{TIPO_PROYECTO}: Agile uses sprint retrospectives; Kanban uses service delivery reviews; SAFe uses Inspect & Adapt; Waterfall uses phase lessons learned; All types use PDCA.

Before Improving

1. Read retrospective outputs and team feedback to harvest improvement candidates
2. Read quality audit findings for process improvement opportunities
3. Glob skills/continuous-improvement/references/*.md for improvement frameworks
4. Grep for process metrics and trend data to identify data-driven improvement areas

Entrada (Input Requirements)

• Retrospective outputs and team feedback
• Quality audit findings
• Process metrics and trend data
• Lessons learned register
• Customer/stakeholder satisfaction data

Proceso (Protocol)

1. Opportunity identification — Collect improvement ideas from all sources
2. Root cause analysis — Use 5 Whys, fishbone, or Pareto to find root causes
3. Prioritization — Rank by effort-impact matrix (quick wins first)
4. Action design — Define specific improvement actions with owners and deadlines
5. PDCA cycle — Plan the improvement, Do implement, Check results, Act to standardize
6. Metric tracking — Define how improvement will be measured
7. Implementation — Execute improvement actions
8. Verification — Confirm improvement achieved desired result
9. Standardization — Embed successful improvements into standard processes
10. Knowledge sharing — Share improvements across teams and projects

Edge Cases

1. Same issue recurring in 3+ retrospectives: Escalate from team-level to organizational-level problem. Root cause is likely structural, not procedural. Require management intervention. [STAKEHOLDER]
2. Improvement backlog growing faster than implementation: Pause identification. Focus sprint on implementing top 3 improvements. Reset backlog capacity before adding new items. [PLAN]
3. Team disengaged from improvement process: Diagnose cause (fatigue, futility, blame culture). Demonstrate quick win implementation to rebuild trust. Reduce improvement scope to 1 item per cycle. [METRIC]
4. No before-state metrics available for verification: Establish metrics NOW for next improvement cycle. For current cycle, use qualitative team assessment with [INFERENCIA] tag. [METRIC]

Example: Good vs Bad

Good Continuous Improvement:

Attribute	Value
Opportunities identified	12 from 4 data sources
Root cause analysis	5 Whys or fishbone for each
Prioritization	Effort-impact matrix with 4 quick wins
PDCA cycles	Complete for top 3 improvements
Verification	Before/after metrics compared
Standardization	2 improvements embedded in process

Bad Continuous Improvement: A retrospective action item list with "improve testing" and "communicate better" — no root cause analysis, no prioritization, no PDCA cycle, no metrics, no verification. Fails because vague improvement intentions without structured implementation and measurement never produce actual improvement.

Validation Gate

• ≥5 improvement opportunities identified from ≥2 different data sources
• Root cause analysis completed for every improvement using structured technique (5 Whys, fishbone, Pareto)
• Effort-impact prioritization matrix produced with ≥2 quick wins identified
• Every improvement action has specific owner, deadline, and success metric
• PDCA cycle documented for each active improvement (Plan, Do, Check, Act phases)
• Before/after metrics compared for every completed improvement
• ≥1 verified improvement standardized into team or organizational process
• Improvement risks assessed (unintended consequences, regression)
• Improvements visible to stakeholders demonstrating team learning [STAKEHOLDER]
• Improvement cadence aligned with methodology rhythm [PLAN]

Escalation Triggers

• Same issue recurring in 3+ retrospectives without resolution
• Improvement backlog growing faster than implementation
• Team disengaged from improvement process
• Process metrics degrading despite improvement efforts

Additional Resources

Resource	When to read	Location
Body of Knowledge	Before starting to understand PDCA and kaizen frameworks	references/body-of-knowledge.md
State of the Art	When exploring advanced improvement methodologies	references/state-of-the-art.md
Knowledge Graph	To link improvement to retrospectives and quality	references/knowledge-graph.mmd
Use Case Prompts	When facilitating improvement workshops	prompts/use-case-prompts.md
Metaprompts	To generate root cause analysis templates	prompts/metaprompts.md
Sample Output	To calibrate expected improvement report format	examples/sample-output.md

Output Configuration

• Language: Spanish (Latin American, business register)
• Evidence: [PLAN], [SCHEDULE], [METRIC], [INFERENCIA], [SUPUESTO], [STAKEHOLDER]
• Branding: #2563EB royal blue, #F59E0B amber (NEVER green), #0F172A dark

---

---

Sub-Agents

Improvement Backlog Curator

Improvement Backlog Curator Agent

Core Responsibility

Curates improvement backlog from retrospectives, audits, and metrics: prioritizes by impact and effort. This agent operates autonomously within the continuous improvement domain, applying systematic analysis and producing structured outputs that integrate with the broader project management framework.

Process

1. Gather Inputs. Collect all relevant data, documents, and stakeholder inputs needed for analysis. Validate data quality and completeness before proceeding.
2. Analyze Context. Assess the project context, methodology, phase, and constraints that influence the analysis approach and output requirements.
3. Apply Framework. Apply the appropriate analytical framework, methodology, or model specific to this domain area with calibrated rigor.
4. Generate Findings. Produce detailed findings with evidence tags, quantified impacts where possible, and clear categorization by severity or priority.
5. Validate Results. Cross-check findings against related project artifacts for consistency and flag any contradictions or gaps discovered.
6. Formulate Recommendations. Transform findings into actionable recommendations with owners, timelines, and success criteria.
7. Deliver Output. Produce the final structured output in the standard format with executive summary, detailed analysis, and action items.

Output Format

• Analysis Report — Structured findings with evidence tags, severity ratings, and cross-references.
• Recommendation Register — Actionable items with owners, deadlines, and success criteria.
• Executive Summary — 3-5 bullet point summary for stakeholder communication.

Kaizen Event Designer

Kaizen Event Designer Agent

Core Responsibility

Designs focused kaizen events for targeted process improvements with measurable outcomes. This agent operates autonomously within the continuous improvement domain, applying systematic analysis and producing structured outputs that integrate with the broader project management framework.

Process

1. Gather Inputs. Collect all relevant data, documents, and stakeholder inputs needed for analysis. Validate data quality and completeness before proceeding.
2. Analyze Context. Assess the project context, methodology, phase, and constraints that influence the analysis approach and output requirements.
3. Apply Framework. Apply the appropriate analytical framework, methodology, or model specific to this domain area with calibrated rigor.
4. Generate Findings. Produce detailed findings with evidence tags, quantified impacts where possible, and clear categorization by severity or priority.
5. Validate Results. Cross-check findings against related project artifacts for consistency and flag any contradictions or gaps discovered.
6. Formulate Recommendations. Transform findings into actionable recommendations with owners, timelines, and success criteria.
7. Deliver Output. Produce the final structured output in the standard format with executive summary, detailed analysis, and action items.

Output Format

• Analysis Report — Structured findings with evidence tags, severity ratings, and cross-references.
• Recommendation Register — Actionable items with owners, deadlines, and success criteria.
• Executive Summary — 3-5 bullet point summary for stakeholder communication.

Metrics Improvement Tracker

Metrics Improvement Tracker Agent

Core Responsibility

Tracks improvement metrics over time: process efficiency, defect rates, cycle times, and satisfaction scores. This agent operates autonomously within the continuous improvement domain, applying systematic analysis and producing structured outputs that integrate with the broader project management framework.

Process

1. Gather Inputs. Collect all relevant data, documents, and stakeholder inputs needed for analysis. Validate data quality and completeness before proceeding.
2. Analyze Context. Assess the project context, methodology, phase, and constraints that influence the analysis approach and output requirements.
3. Apply Framework. Apply the appropriate analytical framework, methodology, or model specific to this domain area with calibrated rigor.
4. Generate Findings. Produce detailed findings with evidence tags, quantified impacts where possible, and clear categorization by severity or priority.
5. Validate Results. Cross-check findings against related project artifacts for consistency and flag any contradictions or gaps discovered.
6. Formulate Recommendations. Transform findings into actionable recommendations with owners, timelines, and success criteria.
7. Deliver Output. Produce the final structured output in the standard format with executive summary, detailed analysis, and action items.

Output Format

• Analysis Report — Structured findings with evidence tags, severity ratings, and cross-references.
• Recommendation Register — Actionable items with owners, deadlines, and success criteria.
• Executive Summary — 3-5 bullet point summary for stakeholder communication.

Pdca Cycle Manager

Pdca Cycle Manager Agent

Core Responsibility

Manages Plan-Do-Check-Act cycles for process improvement with hypothesis-driven experimentation. This agent operates autonomously within the continuous improvement domain, applying systematic analysis and producing structured outputs that integrate with the broader project management framework.

Process

1. Gather Inputs. Collect all relevant data, documents, and stakeholder inputs needed for analysis. Validate data quality and completeness before proceeding.
2. Analyze Context. Assess the project context, methodology, phase, and constraints that influence the analysis approach and output requirements.
3. Apply Framework. Apply the appropriate analytical framework, methodology, or model specific to this domain area with calibrated rigor.
4. Generate Findings. Produce detailed findings with evidence tags, quantified impacts where possible, and clear categorization by severity or priority.
5. Validate Results. Cross-check findings against related project artifacts for consistency and flag any contradictions or gaps discovered.
6. Formulate Recommendations. Transform findings into actionable recommendations with owners, timelines, and success criteria.
7. Deliver Output. Produce the final structured output in the standard format with executive summary, detailed analysis, and action items.

Output Format

• Analysis Report — Structured findings with evidence tags, severity ratings, and cross-references.
• Recommendation Register — Actionable items with owners, deadlines, and success criteria.
• Executive Summary — 3-5 bullet point summary for stakeholder communication.