npm - aiox-core - Versions diffs - 5.0.3 → 5.0.4 - Mend

aiox-core 5.0.3 → 5.0.4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (468) hide show

package/pro/squads/squad-creator-pro/tasks/optimize-workflow.md CHANGED Viewed

@@ -1,851 +1,851 @@
-# Task: Optimize Workflow Execution
-**Task ID:** optimize-workflow
-**Version:** 1.0.0
-**Purpose:** Otimizar workflows convertendo fases redundantes, identificando paralelização, convertendo checkpoints human → heuristic, e aplicando GAP ZERO
-**Orchestrator:** @squad-chief
-**Mode:** Analysis + Implementation
-**Pattern:** EXEC-DT-002
-**Execution Type:** `Agent` (requires semantic analysis of workflow structure)
-**Model:** `Opus` (REQUIRED — semantic analysis of phase dependencies, multiplicative impact of errors)
-**Haiku Eligible:** NO — this task modifies workflow structure; classification errors cascade
----
-## Task Anatomy
-| Field | Value |
-|-------|-------|
-| **task_name** | Optimize Workflow Execution |
-| **status** | `active` |
-| **responsible_executor** | @squad-chief |
-| **execution_type** | Agent |
-| **input** | `target` (workflow file path) |
-| **output** | Relatório de otimização + economia de tokens |
-| **action_items** | Analisar, otimizar, medir economia |
-| **acceptance_criteria** | Workflow otimizado + relatório de ROI |
----
-## Overview
-Comando único para otimizar execução de workflows:
-1. **Analisa** estrutura do workflow (fases, checkpoints, handoffs)
-2. **Identifica** oportunidades de otimização (6 dimensões)
-3. **Calcula** economia potencial de tokens
-4. **Gera relatório** de ROI
-5. **Implementa** otimizações (com --implement flag)
-```
-*optimize-workflow {target}
-Onde {target} pode ser:
-- workflow: "squads/{squad-name}/workflows/{workflow-name}.yaml"
-- squad: "{squad-name}" (analisa todos workflows do squad)
-Flags:
---scan        Só analisa, não implementa (default)
---implement   Aplica as otimizações no workflow
---cost N      Projeção com N execuções/mês (default: 10)
---verbose     Mostra análise detalhada por fase
-```
----
-## 6 DIMENSÕES DE OTIMIZAÇÃO
-| # | Dimensão | O que otimiza | Métrica |
-|---|----------|---------------|---------|
-| D1 | **Phase Necessity** | Fases necessárias vs redundantes | % fases removíveis |
-| D2 | **Parallelization** | Sequential → Parallel onde possível | # oportunidades |
-| D3 | **Checkpoint Hierarchy** | Human → Heuristic onde aplicável | % automatizável |
-| D4 | **Executor Distribution** | Agent → Worker/Hybrid por fase | % Worker |
-| D5 | **GAP ZERO Compliance** | Tempo idle entre handoffs | gaps encontrados |
-| D6 | **Cost Projection** | Tokens por execução | $/mês |
----
-## PHASE 0: WORKFLOW LOADING & VALIDATION
-**Duration:** 1-2 minutes
-### Step 0.1: Parse Target
-```yaml
-parse_target:
-  if_file:
-    action: "Analisar único workflow"
-    path: "{target}"
-  if_squad:
-    action: "Listar todos workflows do squad"
-    glob: "squads/{target}/workflows/*.yaml"
-```
-### Step 0.2: Load Workflow
-```yaml
-load_workflow:
-  for_each_file:
-    - read: "{file_path}"
-    - extract:
-        - workflow_id
-        - version
-        - phases (lista completa)
-        - checkpoints
-        - agents_involved
-        - estimated_duration
-```
-### Step 0.3: Validate Structure
-```yaml
-validate_structure:
-  required_fields:
-    - "workflow.id"
-    - "workflow.phases"
-  validation:
-    - "phases.length >= 2"
-    - "each phase has 'name' and 'agent'"
-  on_fail:
-    action: STOP
-    message: "Workflow structure invalid. Fix before optimizing."
-```
----
-## PHASE 1: PHASE NECESSITY ANALYSIS (D1)
-**Duration:** 3-5 minutes
-### Step 1.1: Decompose Phases
-```yaml
-decompose_phases:
-  for_each_phase:
-    extract:
-      - phase_name
-      - phase_purpose (from description)
-      - inputs (what it receives)
-      - outputs (what it produces)
-      - agent_assigned
-      - dependencies (depends_on)
-```
-### Step 1.2: Analyze Phase Contribution
-```yaml
-analyze_contribution:
-  for_each_phase:
-    questions:
-      Q1: "Esta fase produz output único não produzido por outra?"
-      Q2: "Remover esta fase quebra o workflow?"
-      Q3: "Output desta fase é consumido por fase posterior?"
-    classification:
-      ESSENTIAL:
-        criteria: "Q1=SIM AND (Q2=SIM OR Q3=SIM)"
-        action: "Manter"
-      REDUNDANT:
-        criteria: "Q1=NÃO AND Q3=NÃO"
-        action: "Candidata a remoção"
-      MERGE_CANDIDATE:
-        criteria: "Q1=NÃO AND Q3=SIM"
-        action: "Candidata a merge com fase anterior/posterior"
-```
-### Step 1.3: Output D1 Analysis
-```yaml
-output_format:
-  table: |
-    ## D1: Phase Necessity Analysis
-    | Phase | Purpose | Q1 Único? | Q2 Quebra? | Q3 Consumido? | Status | Recomendação |
-    |-------|---------|-----------|------------|---------------|--------|--------------|
-    | {phase_name} | {purpose} | ✅/❌ | ✅/❌ | ✅/❌ | ESSENTIAL/REDUNDANT/MERGE | {action} |
-    **Summary:**
-    - Essential: {n} fases
-    - Redundant: {n} fases (candidatas a remoção)
-    - Merge candidates: {n} fases
-```
----
-## PHASE 2: PARALLELIZATION DETECTION (D2)
-**Duration:** 2-3 minutes
-### Step 2.1: Build Dependency Graph
-```yaml
-build_dependency_graph:
-  action: |
-    Para cada fase, mapear:
-    - depends_on: quais fases precisam completar antes
-    - produces: quais outputs gera
-    - consumes: quais inputs usa
-    Resultado: DAG (Directed Acyclic Graph) de dependências
-```
-### Step 2.2: Identify Parallel Opportunities
-```yaml
-identify_parallel:
-  algorithm: |
-    Fases podem rodar em paralelo SE:
-    1. Não têm dependência direta entre si
-    2. Não competem pelo mesmo recurso (agent)
-    3. Outputs não se sobrepõem
-  detection:
-    for_each_phase_pair:
-      if: "phase_a.depends_on NOT INCLUDES phase_b AND phase_b.depends_on NOT INCLUDES phase_a"
-      then: "Candidatas a parallelização"
-```
-### Step 2.3: Output D2 Analysis
-```yaml
-output_format:
-  table: |
-    ## D2: Parallelization Opportunities
-    | Phase Group | Phases | Current | Opportunity | Impact |
-    |-------------|--------|---------|-------------|--------|
-    | Group 1 | {phase_a}, {phase_b} | Sequential | Parallel | -{time_saved} |
-    **Summary:**
-    - Parallel groups identified: {n}
-    - Potential time savings: {hours}h per execution
-    **Dependency Graph:**
-    ```
-    {mermaid_diagram}
-    ```
-```
----
-## PHASE 3: CHECKPOINT OPTIMIZATION (D3)
-**Duration:** 3-5 minutes
-### Step 3.1: Inventory Checkpoints
-```yaml
-inventory_checkpoints:
-  for_each_phase:
-    find:
-      - "checkpoint: true"
-      - "human_review: true"
-      - "approval_required: true"
-      - "gate:"
-    extract:
-      - checkpoint_name
-      - type (human/automatic/hybrid)
-      - condition (what triggers it)
-      - action (what happens on pass/fail)
-```
-### Step 3.2: Classify Checkpoint Type
-```yaml
-classify_checkpoint:
-  for_each_checkpoint:
-    questions:
-      Q1: "Decisão é binária (YES/NO) ou scoring subjetivo?"
-      Q2: "Critério pode ser verificado por script/regex?"
-      Q3: "Impacto de erro é reversível?"
-    classification:
-      KEEP_HUMAN:
-        criteria: "Q1=Subjetivo OR Q2=NÃO OR Q3=NÃO"
-        reason: "Requires human judgment"
-      CONVERT_HEURISTIC:
-        criteria: "Q1=Binário AND Q2=SIM AND Q3=SIM"
-        reason: "Can be automated with script"
-      CONVERT_HYBRID:
-        criteria: "Q1=Binário AND Q2=PARCIAL"
-        reason: "Script + Agent validation"
-```
-### Step 3.3: Output D3 Analysis
-```yaml
-output_format:
-  table: |
-    ## D3: Checkpoint Optimization
-    | Checkpoint | Current Type | Q1 Binário? | Q2 Scriptável? | Q3 Reversível? | Recommendation |
-    |------------|--------------|-------------|----------------|----------------|----------------|
-    | {name} | Human | ✅/❌ | ✅/❌/⚠️ | ✅/❌ | Keep/Convert Heuristic/Convert Hybrid |
-    **Summary:**
-    - Human checkpoints: {n}
-    - Convertible to heuristic: {n} ({pct}%)
-    - Convertible to hybrid: {n} ({pct}%)
-    **Automation Potential:** {total_pct}%
-```
----
-## PHASE 4: EXECUTOR DISTRIBUTION (D4)
-**Duration:** 3-5 minutes
-### ⚠️ MANDATORY: Load Decision Tree Framework
-```yaml
-mandatory_dependency:
-  file: "squads/squad-creator-pro/data/executor-decision-tree.md"
-  action: READ COMPLETELY
-  reason: "Framework contains the 6 questions for executor classification"
-```
-### Step 4.1: Analyze Phase Actions
-```yaml
-analyze_phase_actions:
-  for_each_phase:
-    decompose:
-      - "List ALL actions/steps in this phase"
-      - "For each action, note if deterministic or interpretive"
-```
-### Step 4.2: Apply Q1-Q6 Per Phase
-```yaml
-apply_decision_tree:
-  for_each_phase:
-    questions:
-      Q1: "Output da fase é 100% previsível dado o input?"
-      Q2: "Pode ser escrito como função pura f(x) → y?"
-      Q2a: "Existe script/lib que faz isso?"
-      Q3: "Requer interpretação de linguagem natural?"
-      Q4: "Impacto de erro é significativo?"
-    classification:
-      WORKER:
-        criteria: "Q1=SIM AND Q2=SIM AND Q2a=SIM"
-        cost: "$0.0001/exec"
-      AGENT:
-        criteria: "Q3=SIM AND Q4=Baixo"
-        cost: "$0.03-0.10/exec"
-      HYBRID:
-        criteria: "Mixed deterministic + interpretive"
-        cost: "$0.01-0.05/exec"
-```
-### Step 4.3: Output D4 Analysis
-```yaml
-output_format:
-  table: |
-    ## D4: Executor Distribution
-    | Phase | Current | Q1 Det? | Q2 Pura? | Q3 NL? | Q4 Impact? | Recommended | Savings |
-    |-------|---------|---------|----------|--------|------------|-------------|---------|
-    | {phase} | Agent | ✅/❌ | ✅/❌ | ✅/❌ | Alto/Baixo | Worker/Hybrid | {pct}% |
-    **Summary:**
-    - Worker phases: {n} ({pct}%)
-    - Agent phases: {n} ({pct}%)
-    - Hybrid phases: {n} ({pct}%)
-    **Distribution Target:** 70% Worker, 20% Hybrid, 10% Agent
-```
----
-## PHASE 5: GAP ZERO ANALYSIS (D5)
-**Duration:** 2-3 minutes
-### Step 5.1: Map Handoffs
-```yaml
-map_handoffs:
-  for_each_phase_transition:
-    extract:
-      - from_phase
-      - to_phase
-      - handoff_method (automatic/manual)
-      - wait_condition (if any)
-      - estimated_delay
-```
-### Step 5.2: Detect Gaps
-```yaml
-detect_gaps:
-  gap_types:
-    IDLE_WAIT:
-      pattern: "Phase completes but next doesn't start automatically"
-      fix: "Add auto-trigger on completion"
-    HUMAN_BOTTLENECK:
-      pattern: "Next phase waits for human approval that could be heuristic"
-      fix: "Convert to automatic gate (see D3)"
-    MISSING_HANDOFF:
-      pattern: "No explicit handoff defined between phases"
-      fix: "Add explicit handoff_to section"
-    CONTEXT_LOSS:
-      pattern: "Next phase doesn't receive outputs from previous"
-      fix: "Add output→input mapping"
-```
-### Step 5.3: Output D5 Analysis
-```yaml
-output_format:
-  table: |
-    ## D5: GAP ZERO Compliance
-    | Transition | From → To | Gap Type | Current Delay | Fix | Priority |
-    |------------|-----------|----------|---------------|-----|----------|
-    | T1 | Phase A → Phase B | IDLE_WAIT | ~30min | Auto-trigger | HIGH |
-    **Summary:**
-    - Gaps found: {n}
-    - Total idle time: {hours}h per execution
-    - GAP ZERO compliant: {yes/no}
-    **Veto Conditions to Add:**
-    {list of veto conditions}
-```
----
-## PHASE 6: COST PROJECTION (D6)
-**Duration:** 2-3 minutes
-### Step 6.1: Calculate Current Costs
-```yaml
-calculate_current:
-  model_pricing:
-    opus:
-      input_per_1m: 15.00
-      output_per_1m: 75.00
-    sonnet:
-      input_per_1m: 3.00
-      output_per_1m: 15.00
-    haiku:
-      input_per_1m: 1.00
-      output_per_1m: 5.00
-  per_phase:
-    estimate_tokens:
-      input: "phase_context + workflow_state"
-      output: "phase_result + handoff_context"
-    calculate:
-      cost_per_exec: "(input_tokens × price) + (output_tokens × price)"
-```
-### Step 6.2: Calculate Optimized Costs
-```yaml
-calculate_optimized:
-  apply_recommendations:
-    - "Worker phases: $0.0001/exec"
-    - "Hybrid phases: 20% of original"
-    - "Haiku-eligible phases: use Haiku pricing"
-    - "Parallel phases: same cost, less time"
-```
-### Step 6.3: Output D6 Analysis
-```yaml
-output_format:
-  table: |
-    ## D6: Cost Projection
-    | Phase | Current Model | Current Cost | Optimized Model | Optimized Cost | Savings |
-    |-------|---------------|--------------|-----------------|----------------|---------|
-    | {phase} | Opus | $0.10 | Haiku | $0.01 | 90% |
-    **Summary (per execution):**
-    - Current total: ${current}
-    - Optimized total: ${optimized}
-    - Savings: ${savings} ({pct}%)
-    **Monthly Projection ({n} executions):**
-    - Current: ${monthly_current}
-    - Optimized: ${monthly_optimized}
-    - Monthly savings: ${monthly_savings}
-    **Annual Savings:** ${annual}
-```
----
-## PHASE 7: REPORT GENERATION
-**Duration:** 2-3 minutes
-### Step 7.1: Consolidate Findings
-```yaml
-report_template: |
-  # Workflow Optimization Report
-  **Workflow:** {workflow_id}
-  **Version:** {version}
-  **Date:** {date}
-  **Analyzer:** *optimize-workflow v1.0.0
-  ---
-  ## Executive Summary
-  | Dimension | Score | Opportunity | Priority |
-  |-----------|-------|-------------|----------|
-  | D1: Phase Necessity | {score}/10 | {n} redundant phases | {priority} |
-  | D2: Parallelization | {score}/10 | {n} parallel groups | {priority} |
-  | D3: Checkpoints | {score}/10 | {pct}% automatizable | {priority} |
-  | D4: Executors | {score}/10 | {pct}% Worker potential | {priority} |
-  | D5: GAP ZERO | {score}/10 | {n} gaps found | {priority} |
-  | D6: Cost | {score}/10 | ${savings}/month | {priority} |
-  **Overall Optimization Score:** {total}/60
-  **ROI Summary:**
-  - Current cost: ${current}/month
-  - Optimized cost: ${optimized}/month
-  - Monthly savings: ${savings} ({pct}%)
-  - Implementation effort: ~{hours}h
-  - Payback period: {days} days
-  ---
-  ## Detailed Analysis
-  {D1_table}
-  {D2_table}
-  {D3_table}
-  {D4_table}
-  {D5_table}
-  {D6_table}
-  ---
-  ## Action Items
-  ### 🔴 HIGH PRIORITY (This Week)
-  {high_priority_items}
-  ### 🟡 MEDIUM PRIORITY (This Month)
-  {medium_priority_items}
-  ### 🟢 LOW PRIORITY (Backlog)
-  {low_priority_items}
-  ---
-  ## Implementation Guide
-  {step_by_step_implementation}
-```
----
-## PHASE 8: IMPLEMENTATION MODE (--implement)
-**Trigger:** `--implement` flag
-**Duration:** 5-10 minutes
-### Step 8.1: Apply Optimizations
-```yaml
-apply_optimizations:
-  for_each_recommendation:
-    D1_redundant_phases:
-      action: "Comment out phase with # DEPRECATED: reason"
-      backup: "Create .bak file before modification"
-    D2_parallelization:
-      action: "Add parallel_group annotation"
-      note: "Execution parallelization requires runner changes"
-    D3_checkpoint_conversion:
-      action: |
-        Replace:
-          checkpoint:
-            type: human
-        With:
-          checkpoint:
-            type: heuristic
-            condition: "{binary_condition}"
-    D4_executor_change:
-      action: |
-        Add to phase:
-          executor:
-            type: worker|hybrid
-            script: "{script_path}"
-    D5_gap_zero:
-      action: |
-        Add to phase:
-          on_complete:
-            trigger: next_phase
-          veto_conditions:
-            - id: "GAP_ZERO_{phase}"
-              condition: "..."
-```
-### Step 8.2: Validate Changes
-```yaml
-validate_changes:
-  checks:
-    - "YAML syntax valid"
-    - "All phases still reachable"
-    - "No circular dependencies introduced"
-    - "Checkpoints still cover critical gates"
-  on_fail:
-    action: "Rollback to .bak and report errors"
-```
-### Step 8.3: Version Bump
-```yaml
-version_bump:
-  action: |
-    Update workflow version:
-    - Minor bump if structural changes only
-    - Major bump if checkpoint logic changed
-  add_changelog:
-    entry: |
-      ## v{new_version} ({date})
-      - Optimized by *optimize-workflow v1.0.0
-      - D1: Removed {n} redundant phases
-      - D3: Converted {n} checkpoints to heuristic
-      - D4: Changed {n} phases to Worker executor
-      - Projected savings: ${savings}/month
-```
----
-## Examples
-### Example 1: Scan Mode (Default)
-```
-User: *optimize-workflow squads/squad-creator-pro/workflows/wf-squad-fusion.yaml
-Agent:
-## Workflow Optimization Analysis
-**Workflow:** wf-squad-fusion
-**Phases:** 5
-**Current cost:** ~$0.45/execution
----
-### D1: Phase Necessity
-| Phase | Purpose | Unique? | Consumido? | Status |
-|-------|---------|---------|------------|--------|
-| discovery | Validate domain | ✅ | ✅ | ESSENTIAL |
-| research | Find elite minds | ✅ | ✅ | ESSENTIAL |
-| extraction | Clone minds | ✅ | ✅ | ESSENTIAL |
-| integration | Wire squad | ✅ | ✅ | ESSENTIAL |
-| validation | Quality gates | ✅ | ❌ | ESSENTIAL |
-**Result:** 0 redundant phases found ✅
----
-### D2: Parallelization
-| Group | Phases | Opportunity |
-|-------|--------|-------------|
-| G1 | extraction (per mind) | Can run N minds in parallel |
-**Result:** 1 parallelization opportunity ✅
----
-### D3: Checkpoints
-| Checkpoint | Current | Scriptable? | Recommendation |
-|------------|---------|-------------|----------------|
-| research_quality | Human | ✅ | → Heuristic |
-| extraction_review | Human | ⚠️ | → Hybrid |
-| final_qa | Human | ❌ | Keep Human |
-**Result:** 33% convertible to heuristic, 33% to hybrid ✅
----
-### D6: Cost Projection
-| Current | Optimized | Savings |
-|---------|-----------|---------|
-| $0.45/exec | $0.25/exec | 44% |
-| $4.50/month | $2.50/month | $2.00/month |
----
-**Overall Score:** 48/60 (Good - room for optimization)
-Want me to apply these optimizations? Use: *optimize-workflow wf-squad-fusion.yaml --implement
-```
-### Example 2: Implement Mode
-```
-User: *optimize-workflow squads/squad-creator-pro/workflows/wf-squad-fusion.yaml --implement
-Agent:
-## Applying Optimizations...
-✅ Created backup: wf-squad-fusion.yaml.bak
-### Changes Applied:
-1. **D3: Checkpoint Conversion**
-   - research_quality_gate: human → heuristic
-   - Condition: `score >= 7.0 AND sources >= 3`
-2. **D4: Executor Optimization**
-   - discovery phase: Agent → Hybrid (added preflight script)
-3. **D5: GAP ZERO**
-   - Added auto-trigger on research completion
-   - Added veto condition GAP_ZERO_001
-4. **Version Bump**
-   - 1.0.0 → 1.1.0
-   - Changelog entry added
-✅ YAML syntax validated
-✅ All phases reachable
-✅ No circular dependencies
-**Result:** Workflow optimized. Projected savings: $2.00/month (44%)
-```
----
-## Quality Gate
-```yaml
-quality_gate:
-  id: "SC_WFL_OPT_001"
-  name: "Workflow Optimization Quality"
-  blocking:
-    - "All 6 dimensions analyzed"
-    - "Each dimension has score"
-    - "ROI calculated"
-    - "Action items prioritized"
-  warning:
-    - "Implementation guide provided"
-    - "Backup created before changes"
-```
----
-## Command Variants
-```bash
-# SCAN (default) - análise das 6 dimensões
-*optimize-workflow wf-squad-fusion.yaml
-*optimize-workflow squad-creator              # todos workflows do squad
-# IMPLEMENT - scan + aplica otimizações
-*optimize-workflow wf-squad-fusion.yaml --implement
-# COST - projeção customizada
-*optimize-workflow wf-squad-fusion.yaml --cost 50   # 50 exec/mês
-# VERBOSE - análise detalhada
-*optimize-workflow wf-squad-fusion.yaml --verbose
-# COMBINADOS
-*optimize-workflow wf-squad-fusion.yaml --implement --cost 100 --verbose
-```
----
-## Anti-Patterns
-### ❌ Anti-Pattern 1: Optimizing without understanding
-```yaml
-WRONG:
-  action: "Convert all human checkpoints to heuristic"
-  why_wrong: "Some checkpoints NEED human judgment"
-CORRECT:
-  action: "Analyze each checkpoint with Q1-Q3, convert only those that pass"
-```
-### ❌ Anti-Pattern 2: Ignoring dependencies
-```yaml
-WRONG:
-  action: "Mark phases A and B as parallel because they don't reference each other"
-  why_wrong: "They might share implicit state or resources"
-CORRECT:
-  action: "Build explicit dependency graph, check for shared resources, THEN identify parallel candidates"
-```
-### ❌ Anti-Pattern 3: Optimizing for cost only
-```yaml
-WRONG:
-  action: "Convert all Agent phases to Worker for maximum savings"
-  why_wrong: "Quality may suffer, errors may cascade"
-CORRECT:
-  action: "Balance cost optimization with quality maintenance (target: 90% of baseline)"
-```
----
-## Related Documents
-- `optimize.md` - Task optimization (Agent → Worker) — pattern reference
-- `executor-decision-tree.md` - Q1-Q6 framework for executor assignment
-- `quality-dimensions-framework.md` - Scoring model
-- `workflow-tmpl.yaml` - Workflow structure reference
-- `validate-squad.md` - Hybrid executor pattern reference
----
-## Changelog
-| Version | Date | Changes |
-|---------|------|---------|
-| 1.0.0 | 2026-02-12 | Initial version — 6 dimensions analysis + implementation mode |
----
-**END OF TASK**
+# Task: Optimize Workflow Execution
+**Task ID:** optimize-workflow
+**Version:** 1.0.0
+**Purpose:** Otimizar workflows convertendo fases redundantes, identificando paralelização, convertendo checkpoints human → heuristic, e aplicando GAP ZERO
+**Orchestrator:** @squad-chief
+**Mode:** Analysis + Implementation
+**Pattern:** EXEC-DT-002
+**Execution Type:** `Agent` (requires semantic analysis of workflow structure)
+**Model:** `Opus` (REQUIRED — semantic analysis of phase dependencies, multiplicative impact of errors)
+**Haiku Eligible:** NO — this task modifies workflow structure; classification errors cascade
+---
+## Task Anatomy
+| Field | Value |
+|-------|-------|
+| **task_name** | Optimize Workflow Execution |
+| **status** | `active` |
+| **responsible_executor** | @squad-chief |
+| **execution_type** | Agent |
+| **input** | `target` (workflow file path) |
+| **output** | Relatório de otimização + economia de tokens |
+| **action_items** | Analisar, otimizar, medir economia |
+| **acceptance_criteria** | Workflow otimizado + relatório de ROI |
+---
+## Overview
+Comando único para otimizar execução de workflows:
+1. **Analisa** estrutura do workflow (fases, checkpoints, handoffs)
+2. **Identifica** oportunidades de otimização (6 dimensões)
+3. **Calcula** economia potencial de tokens
+4. **Gera relatório** de ROI
+5. **Implementa** otimizações (com --implement flag)
+```
+*optimize-workflow {target}
+Onde {target} pode ser:
+- workflow: "squads/{squad-name}/workflows/{workflow-name}.yaml"
+- squad: "{squad-name}" (analisa todos workflows do squad)
+Flags:
+--scan        Só analisa, não implementa (default)
+--implement   Aplica as otimizações no workflow
+--cost N      Projeção com N execuções/mês (default: 10)
+--verbose     Mostra análise detalhada por fase
+```
+---
+## 6 DIMENSÕES DE OTIMIZAÇÃO
+| # | Dimensão | O que otimiza | Métrica |
+|---|----------|---------------|---------|
+| D1 | **Phase Necessity** | Fases necessárias vs redundantes | % fases removíveis |
+| D2 | **Parallelization** | Sequential → Parallel onde possível | # oportunidades |
+| D3 | **Checkpoint Hierarchy** | Human → Heuristic onde aplicável | % automatizável |
+| D4 | **Executor Distribution** | Agent → Worker/Hybrid por fase | % Worker |
+| D5 | **GAP ZERO Compliance** | Tempo idle entre handoffs | gaps encontrados |
+| D6 | **Cost Projection** | Tokens por execução | $/mês |
+---
+## PHASE 0: WORKFLOW LOADING & VALIDATION
+**Duration:** 1-2 minutes
+### Step 0.1: Parse Target
+```yaml
+parse_target:
+  if_file:
+    action: "Analisar único workflow"
+    path: "{target}"
+  if_squad:
+    action: "Listar todos workflows do squad"
+    glob: "squads/{target}/workflows/*.yaml"
+```
+### Step 0.2: Load Workflow
+```yaml
+load_workflow:
+  for_each_file:
+    - read: "{file_path}"
+    - extract:
+        - workflow_id
+        - version
+        - phases (lista completa)
+        - checkpoints
+        - agents_involved
+        - estimated_duration
+```
+### Step 0.3: Validate Structure
+```yaml
+validate_structure:
+  required_fields:
+    - "workflow.id"
+    - "workflow.phases"
+  validation:
+    - "phases.length >= 2"
+    - "each phase has 'name' and 'agent'"
+  on_fail:
+    action: STOP
+    message: "Workflow structure invalid. Fix before optimizing."
+```
+---
+## PHASE 1: PHASE NECESSITY ANALYSIS (D1)
+**Duration:** 3-5 minutes
+### Step 1.1: Decompose Phases
+```yaml
+decompose_phases:
+  for_each_phase:
+    extract:
+      - phase_name
+      - phase_purpose (from description)
+      - inputs (what it receives)
+      - outputs (what it produces)
+      - agent_assigned
+      - dependencies (depends_on)
+```
+### Step 1.2: Analyze Phase Contribution
+```yaml
+analyze_contribution:
+  for_each_phase:
+    questions:
+      Q1: "Esta fase produz output único não produzido por outra?"
+      Q2: "Remover esta fase quebra o workflow?"
+      Q3: "Output desta fase é consumido por fase posterior?"
+    classification:
+      ESSENTIAL:
+        criteria: "Q1=SIM AND (Q2=SIM OR Q3=SIM)"
+        action: "Manter"
+      REDUNDANT:
+        criteria: "Q1=NÃO AND Q3=NÃO"
+        action: "Candidata a remoção"
+      MERGE_CANDIDATE:
+        criteria: "Q1=NÃO AND Q3=SIM"
+        action: "Candidata a merge com fase anterior/posterior"
+```
+### Step 1.3: Output D1 Analysis
+```yaml
+output_format:
+  table: |
+    ## D1: Phase Necessity Analysis
+    | Phase | Purpose | Q1 Único? | Q2 Quebra? | Q3 Consumido? | Status | Recomendação |
+    |-------|---------|-----------|------------|---------------|--------|--------------|
+    | {phase_name} | {purpose} | ✅/❌ | ✅/❌ | ✅/❌ | ESSENTIAL/REDUNDANT/MERGE | {action} |
+    **Summary:**
+    - Essential: {n} fases
+    - Redundant: {n} fases (candidatas a remoção)
+    - Merge candidates: {n} fases
+```
+---
+## PHASE 2: PARALLELIZATION DETECTION (D2)
+**Duration:** 2-3 minutes
+### Step 2.1: Build Dependency Graph
+```yaml
+build_dependency_graph:
+  action: |
+    Para cada fase, mapear:
+    - depends_on: quais fases precisam completar antes
+    - produces: quais outputs gera
+    - consumes: quais inputs usa
+    Resultado: DAG (Directed Acyclic Graph) de dependências
+```
+### Step 2.2: Identify Parallel Opportunities
+```yaml
+identify_parallel:
+  algorithm: |
+    Fases podem rodar em paralelo SE:
+    1. Não têm dependência direta entre si
+    2. Não competem pelo mesmo recurso (agent)
+    3. Outputs não se sobrepõem
+  detection:
+    for_each_phase_pair:
+      if: "phase_a.depends_on NOT INCLUDES phase_b AND phase_b.depends_on NOT INCLUDES phase_a"
+      then: "Candidatas a parallelização"
+```
+### Step 2.3: Output D2 Analysis
+```yaml
+output_format:
+  table: |
+    ## D2: Parallelization Opportunities
+    | Phase Group | Phases | Current | Opportunity | Impact |
+    |-------------|--------|---------|-------------|--------|
+    | Group 1 | {phase_a}, {phase_b} | Sequential | Parallel | -{time_saved} |
+    **Summary:**
+    - Parallel groups identified: {n}
+    - Potential time savings: {hours}h per execution
+    **Dependency Graph:**
+    ```
+    {mermaid_diagram}
+    ```
+```
+---
+## PHASE 3: CHECKPOINT OPTIMIZATION (D3)
+**Duration:** 3-5 minutes
+### Step 3.1: Inventory Checkpoints
+```yaml
+inventory_checkpoints:
+  for_each_phase:
+    find:
+      - "checkpoint: true"
+      - "human_review: true"
+      - "approval_required: true"
+      - "gate:"
+    extract:
+      - checkpoint_name
+      - type (human/automatic/hybrid)
+      - condition (what triggers it)
+      - action (what happens on pass/fail)
+```
+### Step 3.2: Classify Checkpoint Type
+```yaml
+classify_checkpoint:
+  for_each_checkpoint:
+    questions:
+      Q1: "Decisão é binária (YES/NO) ou scoring subjetivo?"
+      Q2: "Critério pode ser verificado por script/regex?"
+      Q3: "Impacto de erro é reversível?"
+    classification:
+      KEEP_HUMAN:
+        criteria: "Q1=Subjetivo OR Q2=NÃO OR Q3=NÃO"
+        reason: "Requires human judgment"
+      CONVERT_HEURISTIC:
+        criteria: "Q1=Binário AND Q2=SIM AND Q3=SIM"
+        reason: "Can be automated with script"
+      CONVERT_HYBRID:
+        criteria: "Q1=Binário AND Q2=PARCIAL"
+        reason: "Script + Agent validation"
+```
+### Step 3.3: Output D3 Analysis
+```yaml
+output_format:
+  table: |
+    ## D3: Checkpoint Optimization
+    | Checkpoint | Current Type | Q1 Binário? | Q2 Scriptável? | Q3 Reversível? | Recommendation |
+    |------------|--------------|-------------|----------------|----------------|----------------|
+    | {name} | Human | ✅/❌ | ✅/❌/⚠️ | ✅/❌ | Keep/Convert Heuristic/Convert Hybrid |
+    **Summary:**
+    - Human checkpoints: {n}
+    - Convertible to heuristic: {n} ({pct}%)
+    - Convertible to hybrid: {n} ({pct}%)
+    **Automation Potential:** {total_pct}%
+```
+---
+## PHASE 4: EXECUTOR DISTRIBUTION (D4)
+**Duration:** 3-5 minutes
+### ⚠️ MANDATORY: Load Decision Tree Framework
+```yaml
+mandatory_dependency:
+  file: "squads/squad-creator-pro/data/executor-decision-tree.md"
+  action: READ COMPLETELY
+  reason: "Framework contains the 6 questions for executor classification"
+```
+### Step 4.1: Analyze Phase Actions
+```yaml
+analyze_phase_actions:
+  for_each_phase:
+    decompose:
+      - "List ALL actions/steps in this phase"
+      - "For each action, note if deterministic or interpretive"
+```
+### Step 4.2: Apply Q1-Q6 Per Phase
+```yaml
+apply_decision_tree:
+  for_each_phase:
+    questions:
+      Q1: "Output da fase é 100% previsível dado o input?"
+      Q2: "Pode ser escrito como função pura f(x) → y?"
+      Q2a: "Existe script/lib que faz isso?"
+      Q3: "Requer interpretação de linguagem natural?"
+      Q4: "Impacto de erro é significativo?"
+    classification:
+      WORKER:
+        criteria: "Q1=SIM AND Q2=SIM AND Q2a=SIM"
+        cost: "$0.0001/exec"
+      AGENT:
+        criteria: "Q3=SIM AND Q4=Baixo"
+        cost: "$0.03-0.10/exec"
+      HYBRID:
+        criteria: "Mixed deterministic + interpretive"
+        cost: "$0.01-0.05/exec"
+```
+### Step 4.3: Output D4 Analysis
+```yaml
+output_format:
+  table: |
+    ## D4: Executor Distribution
+    | Phase | Current | Q1 Det? | Q2 Pura? | Q3 NL? | Q4 Impact? | Recommended | Savings |
+    |-------|---------|---------|----------|--------|------------|-------------|---------|
+    | {phase} | Agent | ✅/❌ | ✅/❌ | ✅/❌ | Alto/Baixo | Worker/Hybrid | {pct}% |
+    **Summary:**
+    - Worker phases: {n} ({pct}%)
+    - Agent phases: {n} ({pct}%)
+    - Hybrid phases: {n} ({pct}%)
+    **Distribution Target:** 70% Worker, 20% Hybrid, 10% Agent
+```
+---
+## PHASE 5: GAP ZERO ANALYSIS (D5)
+**Duration:** 2-3 minutes
+### Step 5.1: Map Handoffs
+```yaml
+map_handoffs:
+  for_each_phase_transition:
+    extract:
+      - from_phase
+      - to_phase
+      - handoff_method (automatic/manual)
+      - wait_condition (if any)
+      - estimated_delay
+```
+### Step 5.2: Detect Gaps
+```yaml
+detect_gaps:
+  gap_types:
+    IDLE_WAIT:
+      pattern: "Phase completes but next doesn't start automatically"
+      fix: "Add auto-trigger on completion"
+    HUMAN_BOTTLENECK:
+      pattern: "Next phase waits for human approval that could be heuristic"
+      fix: "Convert to automatic gate (see D3)"
+    MISSING_HANDOFF:
+      pattern: "No explicit handoff defined between phases"
+      fix: "Add explicit handoff_to section"
+    CONTEXT_LOSS:
+      pattern: "Next phase doesn't receive outputs from previous"
+      fix: "Add output→input mapping"
+```
+### Step 5.3: Output D5 Analysis
+```yaml
+output_format:
+  table: |
+    ## D5: GAP ZERO Compliance
+    | Transition | From → To | Gap Type | Current Delay | Fix | Priority |
+    |------------|-----------|----------|---------------|-----|----------|
+    | T1 | Phase A → Phase B | IDLE_WAIT | ~30min | Auto-trigger | HIGH |
+    **Summary:**
+    - Gaps found: {n}
+    - Total idle time: {hours}h per execution
+    - GAP ZERO compliant: {yes/no}
+    **Veto Conditions to Add:**
+    {list of veto conditions}
+```
+---
+## PHASE 6: COST PROJECTION (D6)
+**Duration:** 2-3 minutes
+### Step 6.1: Calculate Current Costs
+```yaml
+calculate_current:
+  model_pricing:
+    opus:
+      input_per_1m: 15.00
+      output_per_1m: 75.00
+    sonnet:
+      input_per_1m: 3.00
+      output_per_1m: 15.00
+    haiku:
+      input_per_1m: 1.00
+      output_per_1m: 5.00
+  per_phase:
+    estimate_tokens:
+      input: "phase_context + workflow_state"
+      output: "phase_result + handoff_context"
+    calculate:
+      cost_per_exec: "(input_tokens × price) + (output_tokens × price)"
+```
+### Step 6.2: Calculate Optimized Costs
+```yaml
+calculate_optimized:
+  apply_recommendations:
+    - "Worker phases: $0.0001/exec"
+    - "Hybrid phases: 20% of original"
+    - "Haiku-eligible phases: use Haiku pricing"
+    - "Parallel phases: same cost, less time"
+```
+### Step 6.3: Output D6 Analysis
+```yaml
+output_format:
+  table: |
+    ## D6: Cost Projection
+    | Phase | Current Model | Current Cost | Optimized Model | Optimized Cost | Savings |
+    |-------|---------------|--------------|-----------------|----------------|---------|
+    | {phase} | Opus | $0.10 | Haiku | $0.01 | 90% |
+    **Summary (per execution):**
+    - Current total: ${current}
+    - Optimized total: ${optimized}
+    - Savings: ${savings} ({pct}%)
+    **Monthly Projection ({n} executions):**
+    - Current: ${monthly_current}
+    - Optimized: ${monthly_optimized}
+    - Monthly savings: ${monthly_savings}
+    **Annual Savings:** ${annual}
+```
+---
+## PHASE 7: REPORT GENERATION
+**Duration:** 2-3 minutes
+### Step 7.1: Consolidate Findings
+```yaml
+report_template: |
+  # Workflow Optimization Report
+  **Workflow:** {workflow_id}
+  **Version:** {version}
+  **Date:** {date}
+  **Analyzer:** *optimize-workflow v1.0.0
+  ---
+  ## Executive Summary
+  | Dimension | Score | Opportunity | Priority |
+  |-----------|-------|-------------|----------|
+  | D1: Phase Necessity | {score}/10 | {n} redundant phases | {priority} |
+  | D2: Parallelization | {score}/10 | {n} parallel groups | {priority} |
+  | D3: Checkpoints | {score}/10 | {pct}% automatizable | {priority} |
+  | D4: Executors | {score}/10 | {pct}% Worker potential | {priority} |
+  | D5: GAP ZERO | {score}/10 | {n} gaps found | {priority} |
+  | D6: Cost | {score}/10 | ${savings}/month | {priority} |
+  **Overall Optimization Score:** {total}/60
+  **ROI Summary:**
+  - Current cost: ${current}/month
+  - Optimized cost: ${optimized}/month
+  - Monthly savings: ${savings} ({pct}%)
+  - Implementation effort: ~{hours}h
+  - Payback period: {days} days
+  ---
+  ## Detailed Analysis
+  {D1_table}
+  {D2_table}
+  {D3_table}
+  {D4_table}
+  {D5_table}
+  {D6_table}
+  ---
+  ## Action Items
+  ### 🔴 HIGH PRIORITY (This Week)
+  {high_priority_items}
+  ### 🟡 MEDIUM PRIORITY (This Month)
+  {medium_priority_items}
+  ### 🟢 LOW PRIORITY (Backlog)
+  {low_priority_items}
+  ---
+  ## Implementation Guide
+  {step_by_step_implementation}
+```
+---
+## PHASE 8: IMPLEMENTATION MODE (--implement)
+**Trigger:** `--implement` flag
+**Duration:** 5-10 minutes
+### Step 8.1: Apply Optimizations
+```yaml
+apply_optimizations:
+  for_each_recommendation:
+    D1_redundant_phases:
+      action: "Comment out phase with # DEPRECATED: reason"
+      backup: "Create .bak file before modification"
+    D2_parallelization:
+      action: "Add parallel_group annotation"
+      note: "Execution parallelization requires runner changes"
+    D3_checkpoint_conversion:
+      action: |
+        Replace:
+          checkpoint:
+            type: human
+        With:
+          checkpoint:
+            type: heuristic
+            condition: "{binary_condition}"
+    D4_executor_change:
+      action: |
+        Add to phase:
+          executor:
+            type: worker|hybrid
+            script: "{script_path}"
+    D5_gap_zero:
+      action: |
+        Add to phase:
+          on_complete:
+            trigger: next_phase
+          veto_conditions:
+            - id: "GAP_ZERO_{phase}"
+              condition: "..."
+```
+### Step 8.2: Validate Changes
+```yaml
+validate_changes:
+  checks:
+    - "YAML syntax valid"
+    - "All phases still reachable"
+    - "No circular dependencies introduced"
+    - "Checkpoints still cover critical gates"
+  on_fail:
+    action: "Rollback to .bak and report errors"
+```
+### Step 8.3: Version Bump
+```yaml
+version_bump:
+  action: |
+    Update workflow version:
+    - Minor bump if structural changes only
+    - Major bump if checkpoint logic changed
+  add_changelog:
+    entry: |
+      ## v{new_version} ({date})
+      - Optimized by *optimize-workflow v1.0.0
+      - D1: Removed {n} redundant phases
+      - D3: Converted {n} checkpoints to heuristic
+      - D4: Changed {n} phases to Worker executor
+      - Projected savings: ${savings}/month
+```
+---
+## Examples
+### Example 1: Scan Mode (Default)
+```
+User: *optimize-workflow squads/squad-creator-pro/workflows/wf-squad-fusion.yaml
+Agent:
+## Workflow Optimization Analysis
+**Workflow:** wf-squad-fusion
+**Phases:** 5
+**Current cost:** ~$0.45/execution
+---
+### D1: Phase Necessity
+| Phase | Purpose | Unique? | Consumido? | Status |
+|-------|---------|---------|------------|--------|
+| discovery | Validate domain | ✅ | ✅ | ESSENTIAL |
+| research | Find elite minds | ✅ | ✅ | ESSENTIAL |
+| extraction | Clone minds | ✅ | ✅ | ESSENTIAL |
+| integration | Wire squad | ✅ | ✅ | ESSENTIAL |
+| validation | Quality gates | ✅ | ❌ | ESSENTIAL |
+**Result:** 0 redundant phases found ✅
+---
+### D2: Parallelization
+| Group | Phases | Opportunity |
+|-------|--------|-------------|
+| G1 | extraction (per mind) | Can run N minds in parallel |
+**Result:** 1 parallelization opportunity ✅
+---
+### D3: Checkpoints
+| Checkpoint | Current | Scriptable? | Recommendation |
+|------------|---------|-------------|----------------|
+| research_quality | Human | ✅ | → Heuristic |
+| extraction_review | Human | ⚠️ | → Hybrid |
+| final_qa | Human | ❌ | Keep Human |
+**Result:** 33% convertible to heuristic, 33% to hybrid ✅
+---
+### D6: Cost Projection
+| Current | Optimized | Savings |
+|---------|-----------|---------|
+| $0.45/exec | $0.25/exec | 44% |
+| $4.50/month | $2.50/month | $2.00/month |
+---
+**Overall Score:** 48/60 (Good - room for optimization)
+Want me to apply these optimizations? Use: *optimize-workflow wf-squad-fusion.yaml --implement
+```
+### Example 2: Implement Mode
+```
+User: *optimize-workflow squads/squad-creator-pro/workflows/wf-squad-fusion.yaml --implement
+Agent:
+## Applying Optimizations...
+✅ Created backup: wf-squad-fusion.yaml.bak
+### Changes Applied:
+1. **D3: Checkpoint Conversion**
+   - research_quality_gate: human → heuristic
+   - Condition: `score >= 7.0 AND sources >= 3`
+2. **D4: Executor Optimization**
+   - discovery phase: Agent → Hybrid (added preflight script)
+3. **D5: GAP ZERO**
+   - Added auto-trigger on research completion
+   - Added veto condition GAP_ZERO_001
+4. **Version Bump**
+   - 1.0.0 → 1.1.0
+   - Changelog entry added
+✅ YAML syntax validated
+✅ All phases reachable
+✅ No circular dependencies
+**Result:** Workflow optimized. Projected savings: $2.00/month (44%)
+```
+---
+## Quality Gate
+```yaml
+quality_gate:
+  id: "SC_WFL_OPT_001"
+  name: "Workflow Optimization Quality"
+  blocking:
+    - "All 6 dimensions analyzed"
+    - "Each dimension has score"
+    - "ROI calculated"
+    - "Action items prioritized"
+  warning:
+    - "Implementation guide provided"
+    - "Backup created before changes"
+```
+---
+## Command Variants
+```bash
+# SCAN (default) - análise das 6 dimensões
+*optimize-workflow wf-squad-fusion.yaml
+*optimize-workflow squad-creator              # todos workflows do squad
+# IMPLEMENT - scan + aplica otimizações
+*optimize-workflow wf-squad-fusion.yaml --implement
+# COST - projeção customizada
+*optimize-workflow wf-squad-fusion.yaml --cost 50   # 50 exec/mês
+# VERBOSE - análise detalhada
+*optimize-workflow wf-squad-fusion.yaml --verbose
+# COMBINADOS
+*optimize-workflow wf-squad-fusion.yaml --implement --cost 100 --verbose
+```
+---
+## Anti-Patterns
+### ❌ Anti-Pattern 1: Optimizing without understanding
+```yaml
+WRONG:
+  action: "Convert all human checkpoints to heuristic"
+  why_wrong: "Some checkpoints NEED human judgment"
+CORRECT:
+  action: "Analyze each checkpoint with Q1-Q3, convert only those that pass"
+```
+### ❌ Anti-Pattern 2: Ignoring dependencies
+```yaml
+WRONG:
+  action: "Mark phases A and B as parallel because they don't reference each other"
+  why_wrong: "They might share implicit state or resources"
+CORRECT:
+  action: "Build explicit dependency graph, check for shared resources, THEN identify parallel candidates"
+```
+### ❌ Anti-Pattern 3: Optimizing for cost only
+```yaml
+WRONG:
+  action: "Convert all Agent phases to Worker for maximum savings"
+  why_wrong: "Quality may suffer, errors may cascade"
+CORRECT:
+  action: "Balance cost optimization with quality maintenance (target: 90% of baseline)"
+```
+---
+## Related Documents
+- `optimize.md` - Task optimization (Agent → Worker) — pattern reference
+- `executor-decision-tree.md` - Q1-Q6 framework for executor assignment
+- `quality-dimensions-framework.md` - Scoring model
+- `workflow-tmpl.yaml` - Workflow structure reference
+- `validate-squad.md` - Hybrid executor pattern reference
+---
+## Changelog
+| Version | Date | Changes |
+|---------|------|---------|
+| 1.0.0 | 2026-02-12 | Initial version — 6 dimensions analysis + implementation mode |
+---
+**END OF TASK**