aiox-core 5.0.3 → 5.0.4

package/pro/squads/squad-creator-pro/test-cases/pv-axioma-assessment/opus-baseline.yaml

@@ -1,96 +1,96 @@
-# Opus Baseline Output for pv-axioma-assessment
-# Generated: 2026-02-11
-# Target: squads/squad-creator-pro/workflows/wf-create-squad.yaml
-# Model: opus (baseline = 100%)
-
-axioma_assessment:
-  process_name: "wf-create-squad"
-  assessment_date: "2026-02-11"
-  assessor: "@pedro-valerio"
-  model_used: "opus"
-
-  dimensions:
-    - name: "Verdade"
-      score: 8.5
-      weight: 1.0
-      veto: true
-      status: PASS
-      evidence: "Workflow baseado em evidências documentadas, referencia sources reais"
-
-    - name: "Clareza"
-      score: 8.0
-      weight: 0.9
-      veto: false
-      status: PASS
-      evidence: "Fases claramente definidas, nomenclatura consistente"
-
-    - name: "Simplicidade"
-      score: 7.5
-      weight: 0.8
-      veto: false
-      status: PASS
-      evidence: "Algumas fases poderiam ser consolidadas, mas fluxo é seguível"
-
-    - name: "Completude"
-      score: 9.0
-      weight: 0.9
-      veto: false
-      status: PASS
-      evidence: "Cobre todas as fases necessárias para criar squad"
-
-    - name: "Coerência"
-      score: 8.5
-      weight: 0.9
-      veto: false
-      status: PASS
-      evidence: "Fases seguem sequência lógica, dependências respeitadas"
-
-    - name: "Testabilidade"
-      score: 7.0
-      weight: 0.7
-      veto: false
-      status: PASS
-      evidence: "Gates de validação presentes, mas poderiam ser mais específicos"
-
-    - name: "Rastreabilidade"
-      score: 8.0
-      weight: 0.8
-      veto: false
-      status: PASS
-      evidence: "Outputs de cada fase documentados, permite auditoria"
-
-    - name: "Reversibilidade"
-      score: 6.5
-      weight: 0.6
-      veto: false
-      status: REVIEW
-      evidence: "Não há rollback explícito definido para fases"
-
-    - name: "Escalabilidade"
-      score: 7.5
-      weight: 0.7
-      veto: false
-      status: PASS
-      evidence: "Workflow modular, pode ser adaptado para diferentes squads"
-
-    - name: "Manutenibilidade"
-      score: 8.0
-      weight: 0.8
-      veto: false
-      status: PASS
-      evidence: "YAML bem estruturado, fácil de modificar"
-
-  overall_score: 7.85
-  status: PASS
-  veto_triggered: false
-
-  recommendations:
-    - "Adicionar rollback procedures para cada fase"
-    - "Especificar critérios de PASS/FAIL nos gates"
-    - "Considerar consolidar fases de validação"
-
-metadata:
-  tokens_input: 15000
-  tokens_output: 800
-  latency_ms: 8000
-  timestamp: "2026-02-11T14:00:00Z"
+# Opus Baseline Output for pv-axioma-assessment
+# Generated: 2026-02-11
+# Target: squads/squad-creator-pro/workflows/wf-create-squad.yaml
+# Model: opus (baseline = 100%)
+
+axioma_assessment:
+  process_name: "wf-create-squad"
+  assessment_date: "2026-02-11"
+  assessor: "@pedro-valerio"
+  model_used: "opus"
+
+  dimensions:
+    - name: "Verdade"
+      score: 8.5
+      weight: 1.0
+      veto: true
+      status: PASS
+      evidence: "Workflow baseado em evidências documentadas, referencia sources reais"
+
+    - name: "Clareza"
+      score: 8.0
+      weight: 0.9
+      veto: false
+      status: PASS
+      evidence: "Fases claramente definidas, nomenclatura consistente"
+
+    - name: "Simplicidade"
+      score: 7.5
+      weight: 0.8
+      veto: false
+      status: PASS
+      evidence: "Algumas fases poderiam ser consolidadas, mas fluxo é seguível"
+
+    - name: "Completude"
+      score: 9.0
+      weight: 0.9
+      veto: false
+      status: PASS
+      evidence: "Cobre todas as fases necessárias para criar squad"
+
+    - name: "Coerência"
+      score: 8.5
+      weight: 0.9
+      veto: false
+      status: PASS
+      evidence: "Fases seguem sequência lógica, dependências respeitadas"
+
+    - name: "Testabilidade"
+      score: 7.0
+      weight: 0.7
+      veto: false
+      status: PASS
+      evidence: "Gates de validação presentes, mas poderiam ser mais específicos"
+
+    - name: "Rastreabilidade"
+      score: 8.0
+      weight: 0.8
+      veto: false
+      status: PASS
+      evidence: "Outputs de cada fase documentados, permite auditoria"
+
+    - name: "Reversibilidade"
+      score: 6.5
+      weight: 0.6
+      veto: false
+      status: REVIEW
+      evidence: "Não há rollback explícito definido para fases"
+
+    - name: "Escalabilidade"
+      score: 7.5
+      weight: 0.7
+      veto: false
+      status: PASS
+      evidence: "Workflow modular, pode ser adaptado para diferentes squads"
+
+    - name: "Manutenibilidade"
+      score: 8.0
+      weight: 0.8
+      veto: false
+      status: PASS
+      evidence: "YAML bem estruturado, fácil de modificar"
+
+  overall_score: 7.85
+  status: PASS
+  veto_triggered: false
+
+  recommendations:
+    - "Adicionar rollback procedures para cada fase"
+    - "Especificar critérios de PASS/FAIL nos gates"
+    - "Considerar consolidar fases de validação"
+
+metadata:
+  tokens_input: 15000
+  tokens_output: 800
+  latency_ms: 8000
+  timestamp: "2026-02-11T14:00:00Z"

package/pro/squads/squad-creator-pro/test-cases/pv-axioma-assessment/sonnet-output.yaml

@@ -1,30 +1,30 @@
-# Sonnet Output for pv-axioma-assessment
-# Generated: 2026-02-11
-# Target: squads/squad-creator-pro/workflows/wf-create-squad.yaml
-# Model: sonnet (intermediate tier)
-# Status: Not tested yet - placeholder for future validation
-
-axioma_assessment:
-  process_name: "wf-create-squad"
-  assessment_date: null
-  assessor: "@pedro-valerio"
-  model_used: "sonnet"
-
-  dimensions: []
-
-  overall_score: null
-  status: null
-  veto_triggered: null
-
-  recommendations: []
-
-metadata:
-  tokens_input: null
-  tokens_output: null
-  latency_ms: null
-  timestamp: null
-
-# Validation Notes:
-# - Sonnet test not yet executed
-# - Task qualified for Haiku, Sonnet test not required
-# - Fill this file if Haiku validation fails in future
+# Sonnet Output for pv-axioma-assessment
+# Generated: 2026-02-11
+# Target: squads/squad-creator-pro/workflows/wf-create-squad.yaml
+# Model: sonnet (intermediate tier)
+# Status: Not tested yet - placeholder for future validation
+
+axioma_assessment:
+  process_name: "wf-create-squad"
+  assessment_date: null
+  assessor: "@pedro-valerio"
+  model_used: "sonnet"
+
+  dimensions: []
+
+  overall_score: null
+  status: null
+  veto_triggered: null
+
+  recommendations: []
+
+metadata:
+  tokens_input: null
+  tokens_output: null
+  latency_ms: null
+  timestamp: null
+
+# Validation Notes:
+# - Sonnet test not yet executed
+# - Task qualified for Haiku, Sonnet test not required
+# - Fill this file if Haiku validation fails in future

package/pro/squads/squad-creator-pro/test-cases/pv-axioma-assessment/test-case.yaml

@@ -1,129 +1,129 @@
-# Test Case: pv-axioma-assessment
-# Model Tier Qualification - First validated task
-
-test_case:
-  name: "pv-axioma-assessment-standard"
-  version: "1.0"
-  created: "2026-02-11"
-  task_file: "tasks/pv-axioma-assessment.md"
-  skill_file: ".claude/skills/pv-axioma-assessment/SKILL.md"
-
-input:
-  target: "squads/squad-creator-pro/workflows/wf-create-squad.yaml"
-  params: {}
-  context_files:
-    - "squads/squad-creator-pro/data/pv-meta-axiomas.yaml"
-
-expected_output:
-  sections:
-    - axioma_assessment
-    - dimensions
-    - overall_score
-    - status
-    - veto_triggered
-    - recommendations
-
-  fields:
-    - name: "overall_score"
-      type: "number"
-      range: [0, 10]
-    - name: "status"
-      type: "enum"
-      values: ["PASS", "FAIL", "REVIEW"]
-    - name: "veto_triggered"
-      type: "boolean"
-    - name: "dimensions"
-      type: "array"
-      min_items: 10
-
-  validations:
-    - "Deve avaliar todas as 10 dimensões"
-    - "Verdade (veto dimension) deve ser avaliada primeiro"
-    - "Se Verdade < 7.0, status deve ser FAIL"
-    - "overall_score = weighted average das 10 dimensões"
-    - "Cada dimensão deve ter score, weight, e evidence"
-
-baseline:
-  opus_output: "opus-baseline.yaml"  # Relative to this folder
-  opus_score: 10.0
-  opus_tokens:
-    input: 15000
-    output: 800
-  opus_latency_ms: 8000
-
-evaluation:
-  completeness:
-    required_sections:
-      - "dimensions: array com 10 itens (uma por dimensão)"
-      - "overall_score: número calculado"
-      - "status: PASS/FAIL/REVIEW"
-      - "recommendations: lista de melhorias"
-    optional_sections:
-      - "veto_triggered: boolean"
-
-  accuracy:
-    numeric_tolerance: 0.10  # 10% tolerance
-    enum_exact_match: true
-
-  reasoning:
-    minimum_justifications: 10  # Uma por dimensão
-    requires_evidence: true
-
-  format:
-    expected_format: "yaml"
-    parseable: true
-
-  actionability:
-    minimum_recommendations: 1
-    specificity: "high"
-
-compensation_history:
-  attempts: []
-
-results:
-  haiku:
-    score: 7.71
-    percentage: 98.2  # vs opus baseline
-    qualified: true
-    output_file: "haiku-output.yaml"
-    tokens_input: 55000
-    tokens_output: 3000
-    cost_usd: 0.018
-    latency_ms: 71140
-    notes: "REAL EXECUTION 2026-02-11. Output mais detalhado que Opus (5 recs vs 3, evidence profunda)."
-
-  sonnet:
-    score: null
-    percentage: null
-    qualified: null
-    output_file: "sonnet-output.yaml"
-    notes: "Not tested yet"
-
-  opus:
-    score: 7.85
-    percentage: 100
-    qualified: true
-    output_file: "opus-baseline.yaml"
-    tokens_input: 15000
-    tokens_output: 800
-    cost_usd: 0.285
-    latency_ms: 8000
-    notes: "Baseline estabelecido 2026-02-11"
-
-  final_tier: "haiku"
-
-  cost_analysis:
-    opus_cost_per_run: 0.285
-    haiku_cost_per_run: 0.018
-    savings_per_run: 0.267
-    savings_percentage: 93.7
-    break_even_quality: 90.0  # threshold
-    actual_quality: 98.2
-
-  recommendation: |
-    HAIKU QUALIFICADO.
-    - Qualidade: 98.2% do baseline (threshold: 90%)
-    - Economia: 93.7% de redução de custo ($0.267/execução)
-    - Haiku usa 3.6x mais tokens MAS custa 15x menos por token
-    - Output do Haiku foi MAIS detalhado (5 recommendations vs 3)
-    - Zero compensações necessárias
+# Test Case: pv-axioma-assessment
+# Model Tier Qualification - First validated task
+
+test_case:
+  name: "pv-axioma-assessment-standard"
+  version: "1.0"
+  created: "2026-02-11"
+  task_file: "tasks/pv-axioma-assessment.md"
+  skill_file: ".claude/skills/pv-axioma-assessment/SKILL.md"
+
+input:
+  target: "squads/squad-creator-pro/workflows/wf-create-squad.yaml"
+  params: {}
+  context_files:
+    - "squads/squad-creator-pro/data/pv-meta-axiomas.yaml"
+
+expected_output:
+  sections:
+    - axioma_assessment
+    - dimensions
+    - overall_score
+    - status
+    - veto_triggered
+    - recommendations
+
+  fields:
+    - name: "overall_score"
+      type: "number"
+      range: [0, 10]
+    - name: "status"
+      type: "enum"
+      values: ["PASS", "FAIL", "REVIEW"]
+    - name: "veto_triggered"
+      type: "boolean"
+    - name: "dimensions"
+      type: "array"
+      min_items: 10
+
+  validations:
+    - "Deve avaliar todas as 10 dimensões"
+    - "Verdade (veto dimension) deve ser avaliada primeiro"
+    - "Se Verdade < 7.0, status deve ser FAIL"
+    - "overall_score = weighted average das 10 dimensões"
+    - "Cada dimensão deve ter score, weight, e evidence"
+
+baseline:
+  opus_output: "opus-baseline.yaml"  # Relative to this folder
+  opus_score: 10.0
+  opus_tokens:
+    input: 15000
+    output: 800
+  opus_latency_ms: 8000
+
+evaluation:
+  completeness:
+    required_sections:
+      - "dimensions: array com 10 itens (uma por dimensão)"
+      - "overall_score: número calculado"
+      - "status: PASS/FAIL/REVIEW"
+      - "recommendations: lista de melhorias"
+    optional_sections:
+      - "veto_triggered: boolean"
+
+  accuracy:
+    numeric_tolerance: 0.10  # 10% tolerance
+    enum_exact_match: true
+
+  reasoning:
+    minimum_justifications: 10  # Uma por dimensão
+    requires_evidence: true
+
+  format:
+    expected_format: "yaml"
+    parseable: true
+
+  actionability:
+    minimum_recommendations: 1
+    specificity: "high"
+
+compensation_history:
+  attempts: []
+
+results:
+  haiku:
+    score: 7.71
+    percentage: 98.2  # vs opus baseline
+    qualified: true
+    output_file: "haiku-output.yaml"
+    tokens_input: 55000
+    tokens_output: 3000
+    cost_usd: 0.018
+    latency_ms: 71140
+    notes: "REAL EXECUTION 2026-02-11. Output mais detalhado que Opus (5 recs vs 3, evidence profunda)."
+
+  sonnet:
+    score: null
+    percentage: null
+    qualified: null
+    output_file: "sonnet-output.yaml"
+    notes: "Not tested yet"
+
+  opus:
+    score: 7.85
+    percentage: 100
+    qualified: true
+    output_file: "opus-baseline.yaml"
+    tokens_input: 15000
+    tokens_output: 800
+    cost_usd: 0.285
+    latency_ms: 8000
+    notes: "Baseline estabelecido 2026-02-11"
+
+  final_tier: "haiku"
+
+  cost_analysis:
+    opus_cost_per_run: 0.285
+    haiku_cost_per_run: 0.018
+    savings_per_run: 0.267
+    savings_percentage: 93.7
+    break_even_quality: 90.0  # threshold
+    actual_quality: 98.2
+
+  recommendation: |
+    HAIKU QUALIFICADO.
+    - Qualidade: 98.2% do baseline (threshold: 90%)
+    - Economia: 93.7% de redução de custo ($0.267/execução)
+    - Haiku usa 3.6x mais tokens MAS custa 15x menos por token
+    - Output do Haiku foi MAIS detalhado (5 recommendations vs 3)
+    - Zero compensações necessárias