aiox-core 5.0.4 → 5.0.5

package/pro/squads/squad-creator-pro/docs/QUICK-START.md

@@ -1,205 +0,0 @@
-# Quick Start: Resumo Rápido
-
-> **Este é um RESUMO.** Para tutorial completo com exemplos, veja [TUTORIAL-COMPLETO.md](./TUTORIAL-COMPLETO.md).
->
-> **Primeira vez?** Comece por [POR-ONDE-COMECAR.md](./POR-ONDE-COMECAR.md).
-
----
-
-## Pré-requisitos
-
-- Claude Code ativo
-- Projeto AIOS configurado
-
----
-
-## Passo 1: Ativar o Squad Architect (30 seg)
-
-```
-@squad-creator
-```
-
-Você verá o greeting do Squad Architect. Ele está pronto para receber comandos.
-
----
-
-## Passo 2: Solicitar Criação do Squad (1 min)
-
-Simplesmente diga o que você quer:
-
-```
-Quero um squad de copywriting
-```
-
-**O que acontece:**
-1. Squad Architect inicia automaticamente a pesquisa
-2. NÃO pergunta "quer pesquisar?" - já vai direto
-3. Executa 3-5 iterações de pesquisa com devil's advocate
-
----
-
-## Passo 3: Pre-Flight (1 min)
-
-O sistema pergunta sobre o modo de execução:
-
-```
-┌─────────────────────────────────────────────────────────────────┐
-│ PRE-FLIGHT: CRIAÇÃO DE SQUAD                                    │
-├─────────────────────────────────────────────────────────────────┤
-│                                                                 │
-│ Como você quer executar?                                        │
-│                                                                 │
-│ 🚀 YOLO    - Não tenho materiais (60-75% fidelity)             │
-│ 💎 QUALITY - Tenho livros/PDFs (85-95% fidelity)               │
-│ 🔀 HYBRID  - Tenho materiais de alguns experts                  │
-│                                                                 │
-└─────────────────────────────────────────────────────────────────┘
-```
-
-**Escolha:**
-- **YOLO** se quer rapidez e não tem materiais
-- **QUALITY** se tem livros/PDFs dos experts
-- **HYBRID** se tem materiais de alguns, não de todos
-
----
-
-## Passo 4: Aprovar Elite Minds (1 min)
-
-Após a pesquisa, você vê os resultados:
-
-```
-┌─────────────────────────────────────────────────────────────────┐
-│ ELITE MINDS ENCONTRADOS                                         │
-├──────────────────┬──────────┬───────────────────────────────────┤
-│ Mind             │ Tier     │ Framework Principal               │
-├──────────────────┼──────────┼───────────────────────────────────┤
-│ {expert-1}       │ Tier 0   │ {framework-1}                     │
-│ {expert-2}       │ Tier 1   │ {framework-2}                     │
-│ {expert-3}       │ Tier 1   │ {framework-3}                     │
-│ ...              │ ...      │ ...                               │
-└──────────────────┴──────────┴───────────────────────────────────┘
-
-Criar agentes baseados nestes minds?
-```
-
-Responda **sim** para prosseguir.
-
----
-
-## Passo 5: Aguardar Criação (2-3 min)
-
-O sistema executa automaticamente:
-
-```
-✓ Coletando fontes para {expert-name}...
-✓ Extraindo Voice DNA...
-✓ Extraindo Thinking DNA...
-✓ Rodando Smoke Tests (3/3)...
-✓ Criando agent.md...
-
-[Repetir para cada mind]
-
-✓ Criando orchestrator...
-✓ Gerando workflows...
-✓ Validando squad...
-✓ Gerando Quality Dashboard...
-```
-
----
-
-## Passo 6: Receber o Squad Pronto
-
-```
-┌─────────────────────────────────────────────────────────────────┐
-│ ✅ SQUAD CRIADO: {squad-name}                                   │
-├─────────────────────────────────────────────────────────────────┤
-│                                                                 │
-│ Agents: N (experts + 1 orchestrator)                            │
-│ Quality Score: X.X/10                                           │
-│ Fidelity Estimate: XX%                                          │
-│                                                                 │
-│ Ativação: @{squad-name}                                         │
-│ Dashboard: squads/{squad-name}/docs/quality-dashboard.md        │
-│                                                                 │
-└─────────────────────────────────────────────────────────────────┘
-```
-
----
-
-## Usar o Squad
-
-```
-@{squad-name}
-```
-
-Agora você tem acesso a todos os agents do squad:
-
-```
-@{squad-name}:{expert-1}  → Especialidade do expert 1
-@{squad-name}:{expert-2}  → Especialidade do expert 2
-@{squad-name}:{expert-3}  → Especialidade do expert 3
-```
-
----
-
-## Resumo do Fluxo
-
-```
-@squad-creator
-      ↓
-"Quero um squad de {domínio}"
-      ↓
-Escolher modo (YOLO/QUALITY/HYBRID)
-      ↓
-Aprovar elite minds
-      ↓
-Aguardar criação automática
-      ↓
-@{squad} para usar
-```
-
----
-
-## Comandos Úteis
-
-| Comando | O que faz |
-|---------|-----------|
-| `*create-squad` | Criar novo squad (com workflow guiado) |
-| `*clone-mind {name}` | Clonar um expert específico |
-| `*validate-squad {name}` | Validar squad existente |
-| `*quality-dashboard {name}` | Gerar dashboard de qualidade |
-| `*help` | Ver todos os comandos |
-
----
-
-## Próximos Passos
-
-1. **Entender os conceitos:** Leia [CONCEPTS.md](./CONCEPTS.md)
-2. **Ver diagramas:** Leia [ARCHITECTURE-DIAGRAMS.md](./ARCHITECTURE-DIAGRAMS.md)
-3. **Problemas?** Leia [TROUBLESHOOTING.md](./TROUBLESHOOTING.md)
-
----
-
-## Exemplo Real: 30 Segundos
-
-```
-User: @squad-creator
-Bot:  🎨 Squad Architect ready...
-
-User: Quero um squad de {domínio}
-Bot:  I'll research the best minds in {domínio}. Starting...
-      [pesquisa 3-5 iterações]
-
-Bot:  Found N elite minds. Create agents?
-User: sim
-
-Bot:  [cria automaticamente]
-      ✅ Squad {squad-name} criado! Ative com @{squad-name}
-```
-
-**É isso.** Sem configuração, sem setup, sem complicação.
-
----
-
-**Squad Architect | Quick Start v1.0**
-*"De zero a squad em 5 minutos."*

package/pro/squads/squad-creator-pro/docs/RFC-001-deterministic-refactoring.md

@@ -1,463 +0,0 @@
-# RFC-001: Deterministic Operations Refactoring
-
-**Status:** DRAFT
-**Created:** 2026-02-01
-**Author:** Squad Architect
-**Priority:** High
-
----
-
-## Summary
-
-Analysis of squad-creator system revealed **107 deterministic operations** that can be handled by Python scripts, and **38 LLM-required operations** that need semantic understanding.
-
-This RFC proposes separating these concerns for:
-- Faster validation (no LLM latency)
-- Lower cost (no token usage for file checks)
-- More reliable results (deterministic = reproducible)
-- Better error messages (precise line numbers, exact issues)
-
----
-
-## Current State
-
-```
-User: *validate-squad copy
-       ↓
-[LLM does EVERYTHING]
-- Counts files (could be script)
-- Checks YAML syntax (could be script)
-- Verifies references exist (could be script)
-- Judges content quality (needs LLM)
-- Calculates scores (could be script)
-       ↓
-Result: Mixed deterministic + semantic in one pass
-```
-
-**Problems:**
-1. Slow: LLM processes everything sequentially
-2. Expensive: Tokens spent on file counting
-3. Inconsistent: LLM might miss structural issues
-4. No caching: Same checks repeated every time
-
----
-
-## Proposed Architecture
-
-```
-User: *validate-squad copy
-       ↓
-[PHASE 1: DETERMINISTIC SCRIPTS]
-├── script_inventory.py          → Counts files, lists components
-├── script_yaml_validator.py     → Checks YAML syntax, required keys
-├── script_dependency_check.py   → Verifies all references exist
-├── script_naming_validator.py   → Checks kebab-case, camelCase
-└── script_quality_gate.py       → Checks counts, line numbers
-       ↓
-[GATE: Any BLOCKING issues?]
-├── YES → Stop, report issues, no LLM needed
-└── NO  → Continue to Phase 2
-       ↓
-[PHASE 2: LLM SEMANTIC ANALYSIS]
-├── Content quality judgment
-├── Voice consistency check
-├── Example realism assessment
-├── Domain specificity analysis
-└── Architecture coherence
-       ↓
-[PHASE 3: SCORING]
-├── script_scoring.py            → Calculate weighted scores
-└── Compare to thresholds
-       ↓
-Result: Structured report with deterministic + semantic
-```
-
----
-
-## Proposed Scripts
-
-### 1. script_inventory.py
-
-**Operations Handled:** 30+
-
-```python
-# Input
-squad_path: str  # "squads/{squad-name}/"
-
-# Output
-{
-  "exists": true,
-  "components": {
-    "agents": {"count": 22, "files": ["{agent-name}.md", ...]},
-    "tasks": {"count": 58, "files": [...]},
-    "workflows": {"count": 3, "files": [...]},
-    "templates": {"count": 10, "files": [...]},
-    "checklists": {"count": 5, "files": [...]},
-    "data": {"count": 2, "files": [...]}
-  },
-  "required_files": {
-    "config.yaml": true,
-    "README.md": true,
-    "CHANGELOG.md": false
-  }
-}
-```
-
-**What it does:**
-- `os.path.exists()` for directory checks
-- `glob.glob()` for file counting
-- `pathlib` for path operations
-
-### 2. script_yaml_validator.py
-
-**Operations Handled:** 25+
-
-```python
-# Input
-file_path: str  # "squads/{squad-name}/agents/{agent-name}.md"
-expected_keys: list  # ["agent", "persona", "commands", "voice_dna"]
-
-# Output
-{
-  "valid_syntax": true,
-  "parse_errors": [],
-  "missing_keys": ["voice_dna.emotional_states"],
-  "present_keys": ["agent", "persona", "commands", "voice_dna.vocabulary"],
-  "nested_counts": {
-    "voice_dna.vocabulary.always_use": 8,
-    "voice_dna.vocabulary.never_use": 6,
-    "output_examples": 3,
-    "objection_algorithms": 4,
-    "anti_patterns.never_do": 13
-  }
-}
-```
-
-**What it does:**
-- `yaml.safe_load()` for parsing
-- Recursive key checking
-- Array length counting
-
-### 3. script_dependency_check.py
-
-**Operations Handled:** 20+
-
-```python
-# Input
-agent_file: str  # "squads/{squad-name}/agents/{agent-name}.md"
-
-# Output
-{
-  "references": {
-    "tasks": ["create-sales-page.md", "write-headlines.md"],
-    "templates": ["sales-page-tmpl.yaml"],
-    "checklists": ["copy-quality.md"],
-    "data": ["copywriting-kb.md"]
-  },
-  "resolved": {
-    "tasks/create-sales-page.md": true,
-    "tasks/write-headlines.md": true,
-    "templates/sales-page-tmpl.yaml": false,  # MISSING
-    "checklists/copy-quality.md": true,
-    "data/copywriting-kb.md": true
-  },
-  "broken_references": ["templates/sales-page-tmpl.yaml"],
-  "orphan_files": ["tasks/unused-task.md"]
-}
-```
-
-**What it does:**
-- Parse YAML `dependencies` block
-- Check each referenced file exists
-- Scan for files not referenced by any agent
-
-### 4. script_naming_validator.py
-
-**Operations Handled:** 8+
-
-```python
-# Input
-squad_path: str
-
-# Output
-{
-  "config_name_matches_folder": true,
-  "slash_prefix_valid": true,  # camelCase
-  "file_naming": {
-    "agents": {"valid": ["{agent-name}.md"], "invalid": []},
-    "tasks": {"valid": [...], "invalid": ["Create_Sales_Page.md"]}  # Wrong case
-  },
-  "violations": [
-    {"file": "tasks/Create_Sales_Page.md", "issue": "Must be kebab-case"}
-  ]
-}
-```
-
-**What it does:**
-- Regex patterns: `^[a-z0-9]+(-[a-z0-9]+)*\.md$` for kebab-case
-- Regex patterns: `^[a-z][a-zA-Z0-9]*$` for camelCase
-- String comparison for config name vs folder
-
-### 5. script_quality_gate.py
-
-**Operations Handled:** 15+
-
-```python
-# Input
-agent_file: str
-thresholds: dict  # {"min_lines": 300, "min_examples": 3, ...}
-
-# Output
-{
-  "file": "agents/{agent-name}.md",
-  "metrics": {
-    "lines": 450,
-    "vocabulary_always_use": 8,
-    "vocabulary_never_use": 6,
-    "output_examples": 3,
-    "objection_algorithms": 4,
-    "anti_patterns_never_do": 13,
-    "anti_patterns_always_do": 9
-  },
-  "gates": {
-    "lines_check": {"threshold": 300, "value": 450, "pass": true},
-    "examples_check": {"threshold": 3, "value": 3, "pass": true},
-    "vocabulary_check": {"threshold": 5, "value": 8, "pass": true}
-  },
-  "blocking_issues": [],
-  "warnings": []
-}
-```
-
-**What it does:**
-- Count lines: `len(file.readlines())`
-- Count YAML array items
-- Compare against thresholds
-
-### 6. script_scoring.py
-
-**Operations Handled:** 10+
-
-```python
-# Input
-dimension_scores: dict  # From LLM
-{
-  "template_conformance": 9,
-  "principle_adherence": 8,
-  "internal_consistency": 9,
-  "integration_quality": 10,
-  "practical_utility": 9
-}
-weights: dict
-{
-  "template_conformance": 0.25,
-  "principle_adherence": 0.25,
-  "internal_consistency": 0.20,
-  "integration_quality": 0.15,
-  "practical_utility": 0.15
-}
-
-# Output
-{
-  "dimension_scores": {...},
-  "weighted_scores": {
-    "template_conformance": 2.25,  # 9 * 0.25
-    "principle_adherence": 2.00,
-    ...
-  },
-  "total_score": 8.95,
-  "interpretation": "Good - Ready for production",
-  "pass": true,
-  "threshold": 7.0
-}
-```
-
-**What it does:**
-- Multiply scores by weights
-- Sum weighted scores
-- Compare to thresholds
-- Return interpretation
-
-### 7. script_registry_manager.py
-
-**Operations Handled:** 5+
-
-```python
-# Input
-query: str  # "copywriting"
-registry_path: str  # "data/squad-registry.yaml"
-
-# Output
-{
-  "query": "copywriting",
-  "matches": [
-    {
-      "squad": "copy",
-      "domain": "copywriting",
-      "match_type": "domain_index",
-      "confidence": "exact"
-    }
-  ],
-  "related": [
-    {"squad": "creator-os", "reason": "content creation"}
-  ]
-}
-```
-
-**What it does:**
-- Parse squad-registry.yaml
-- Search domain_index
-- Search keywords arrays
-- Return matches with confidence
-
----
-
-## Operation Classification Summary
-
-### DETERMINISTIC (107 operations)
-
-| Category | Count | Script |
-|----------|-------|--------|
-| File/directory existence | 30 | script_inventory.py |
-| YAML structure validation | 25 | script_yaml_validator.py |
-| Reference/dependency checks | 20 | script_dependency_check.py |
-| Naming convention validation | 8 | script_naming_validator.py |
-| Threshold comparisons | 15 | script_quality_gate.py |
-| Score calculation | 10 | script_scoring.py |
-
-### LLM-REQUIRED (38 operations)
-
-| Category | Count | Why LLM Needed |
-|----------|-------|----------------|
-| Content quality judgment | 12 | Semantic understanding |
-| Voice/persona consistency | 6 | Style analysis |
-| Example realism assessment | 5 | Domain knowledge |
-| Architecture design | 5 | Creative judgment |
-| Domain viability | 4 | Business judgment |
-| Gap analysis | 3 | Problem identification |
-| Tier classification | 3 | Expert judgment |
-
----
-
-## Implementation Plan
-
-### Phase 1: Core Scripts (Week 1)
-
-1. **script_inventory.py** - Already created as `refresh-registry.py`
-2. **script_yaml_validator.py** - New
-3. **script_quality_gate.py** - New
-
-### Phase 2: Dependency & Naming (Week 2)
-
-4. **script_dependency_check.py** - New
-5. **script_naming_validator.py** - New
-
-### Phase 3: Scoring & Registry (Week 3)
-
-6. **script_scoring.py** - New
-7. **script_registry_manager.py** - Extend refresh-registry.py
-
-### Phase 4: Integration (Week 4)
-
-8. **script_orchestrator.py** - Chains all scripts
-9. Update tasks to use scripts
-10. Update agent to call scripts first
-
----
-
-## Interface Between Scripts and LLM
-
-### Script Output → LLM Input
-
-```yaml
-# Scripts produce structured data
-deterministic_results:
-  inventory:
-    agents: 22
-    tasks: 58
-  structure_validation:
-    valid: true
-    missing_keys: []
-  dependency_check:
-    broken_references: []
-  quality_gates:
-    all_pass: true
-
-# LLM receives this context
-llm_prompt: |
-  The following structural checks have PASSED:
-  - 22 agents, 58 tasks found
-  - All YAML syntax valid
-  - All dependencies resolved
-  - All naming conventions followed
-
-  Now evaluate SEMANTIC quality:
-  1. Are the examples realistic?
-  2. Is the voice consistent?
-  3. Are anti-patterns domain-specific?
-```
-
-### LLM Output → Script Input
-
-```yaml
-# LLM produces dimension scores
-llm_output:
-  dimension_scores:
-    template_conformance: 9
-    principle_adherence: 8
-    internal_consistency: 9
-    integration_quality: 10
-    practical_utility: 9
-
-# Script calculates final score
-script_scoring.py:
-  input: llm_output.dimension_scores
-  output:
-    total_score: 8.95
-    interpretation: "Good"
-    pass: true
-```
-
----
-
-## Benefits
-
-| Metric | Before | After |
-|--------|--------|-------|
-| Validation time | ~30s (all LLM) | ~5s (scripts) + ~10s (LLM) |
-| Token cost | High (counting files) | Low (only semantic) |
-| Reliability | Variable | Deterministic base |
-| Error messages | Generic | Precise (line numbers) |
-| Caching | None | Script results cached |
-
----
-
-## Risks & Mitigations
-
-| Risk | Mitigation |
-|------|------------|
-| Script bugs give wrong results | Comprehensive unit tests |
-| Scripts and LLM disagree | LLM can override with justification |
-| Maintenance burden | Single source of truth (scripts), clear interfaces |
-| Over-engineering | Start with 3 core scripts, expand as needed |
-
----
-
-## Next Steps
-
-1. **Review this RFC** - Get feedback on approach
-2. **Prototype 3 core scripts** - inventory, yaml_validator, quality_gate
-3. **Test on squad-creator itself** - Validate approach works
-4. **Integrate with tasks** - Update validate-squad.md to use scripts
-5. **Document interfaces** - Clear contracts between scripts and LLM
-
----
-
-## Appendix: Full Operation List
-
-See analysis output for complete list of 145 operations with classifications.
-
----
-
-_RFC-001 Version: 1.0_
-_Status: Awaiting Review_