aiox-core 5.0.4 → 5.0.7

package/pro/squads/squad-creator-pro/tasks/deep-research-validate.md

@@ -0,0 +1,192 @@
+<!-- SINKRA_TASK_METADATA:START -->
+```yaml
+sinkra_task_metadata:
+  task_id: deep-research-validate
+  task_name: Validate Deep Research Output
+  status: pending
+  responsible_executor: '@squad-chief'
+  execution_type: Agent
+  estimated_time: 30m
+  domain: Operational
+  input:
+  - Consultar a seção de inputs no corpo da task
+  output:
+  - Consultar a seção de outputs no corpo da task
+  action_items:
+  - '.1: Check Minimum Requirements'
+  - '.2: Calculate Quality Score'
+  - '.3: Determine Verdict'
+  - '.4: Handle Failure (Retry Logic)'
+  acceptance_criteria:
+  - All 4 minimum requirements checked [threshold: >= 4]
+  - Quality score calculated with correct weights (30/25/25/20) [threshold: >= 60%]
+  - Verdict correctly determined against thresholds [threshold: >= 1 verdict]
+  - 'On FAIL with retries remaining: Phase 3 re-triggered with adjusted queries [threshold: <= 2 retries]'
+  - 'On FAIL after max retries: partial research returned with explicit warning [threshold: >= 1 warning]'
+  output_persistence: transient_output
+  accountable_id: Human:Squad_Operator
+  accountability_scope: review_only
+  escalation_priority: medium
+```
+<!-- SINKRA_TASK_METADATA:END -->
+
+<!-- SINKRA_CONTRACT:START -->
+```yaml
+sinkra_contract:
+  Domain: Tactical
+  atomic_layer: Atom
+  executor: Agent
+  pre_condition: "inputs, dependências e artefatos prévios resolvidos antes de iniciar a execução."
+  post_condition: "output principal gerado, validado e pronto para handoff da próxima fase."
+  performance: "executar dentro do SLA declarado, registrar erro explicitamente e escalar via handoff sem falha silenciosa."
+```
+<!-- SINKRA_CONTRACT:END -->
+
+
+# Task: Validate Deep Research Output
+
+## Task Anatomy
+
+| Field | Value |
+|-------|-------|
+| **Task ID** | `deep-research-validate` |
+| **Version** | `1.0.0` |
+| **Status** | `active` |
+| **Responsible Executor** | `squad-chief` |
+| **Execution Type** | `agent` |
+
+## Metadata
+
+| Key | Value |
+|-----|-------|
+| **Model** | Opus |
+| **Haiku Eligible** | NO |
+| **Parent Task** | `deep-research-pre-agent` (stub) |
+| **Phase** | 4 of 4 |
+| **Created** | 2026-03-26 |
+| **Part of** | squads/squad-creator-pro |
+
+## Purpose
+
+Validate the research document produced in Phase 3 against minimum quality thresholds. Score the output, determine pass/conditional/fail, and trigger retry if below threshold.
+
+## Prerequisites
+
+- Research document from Phase 3 at `docs/research/{slug}-{activity}-research.md`
+- Research prompt from Phase 2 (to verify scope coverage)
+
+## Inputs
+
+| Parameter | Type | Required | Description |
+|-----------|------|----------|-------------|
+| `research_document_path` | string | Yes | Path to the research document from Phase 3 |
+| `research_prompt` | structured text | Yes | The 7-component prompt (for scope coverage check) |
+| `retry_count` | integer | No | Current retry attempt (0-2), default 0 |
+
+## Workflow / Steps
+
+### Step 4.1: Check Minimum Requirements
+
+Verify the research document meets baseline thresholds:
+
+```yaml
+minimum_requirements:
+  total_lines: 500
+  primary_sources: 3
+  scope_coverage: 4/6 sections
+  actionable_content: true  # Has processes, not just theory
+```
+
+### Step 4.2: Calculate Quality Score
+
+Score the document across 4 weighted dimensions:
+
+```yaml
+quality_score:
+  primary_evidence: 30%  # Quotes, real examples
+  scope_coverage: 25%    # Coverage of prompt scope sections
+  actionable_processes: 25%  # Workflows, templates, checklists (not just theory)
+  source_credibility: 20%   # Quality and variety of sources cited
+```
+
+### Step 4.3: Determine Verdict
+
+```yaml
+thresholds:
+  pass: ">= 80%"
+  conditional: "60-79%"
+  fail: "< 60%"
+```
+
+| Verdict | Score | Action |
+|---------|-------|--------|
+| PASS | >= 80% | Accept research document, proceed to agent creation |
+| CONDITIONAL | 60-79% | Accept with warning, flag weak areas for manual enrichment |
+| FAIL | < 60% | Trigger retry (if retries remaining) or accept partial with warning |
+
+### Step 4.4: Handle Failure (Retry Logic)
+
+```yaml
+if quality_score < 60:
+  retry_count += 1
+  if retry_count <= 2:
+    - Adjust queries (broader or more specific based on gaps)
+    - Re-execute Phase 3 (deep-research-execute) with retry_count
+  else:
+    - Return partial research with warning
+    - Flag for manual enrichment
+    - Log: "Research quality below threshold after 2 retries"
+```
+
+## Output
+
+| Artifact | Format | Description |
+|----------|--------|-------------|
+| `validation_result` | YAML | Verdict (PASS/CONDITIONAL/FAIL), score breakdown, gaps |
+| Updated research metadata | Appended to research doc | Quality score added to Research Metadata section |
+
+Example validation result:
+
+```yaml
+validation_result:
+  verdict: "PASS"
+  quality_score: 85
+  breakdown:
+    primary_evidence: 28/30
+    scope_coverage: 20/25
+    actionable_processes: 22/25
+    source_credibility: 15/20
+  minimum_requirements:
+    total_lines: 1247  # >= 500 OK
+    primary_sources: 5  # >= 3 OK
+    scope_coverage: "5/6"  # >= 4/6 OK
+    actionable_content: true
+  gaps: []
+  retry_count: 0
+```
+
+## Acceptance Criteria
+
+- [ ] All 4 minimum requirements checked [threshold: >= 4]
+- [ ] Quality score calculated with correct weights (30/25/25/20) [threshold: >= 60%]
+- [ ] Verdict correctly determined against thresholds [threshold: >= 1 verdict]
+- [ ] On FAIL with retries remaining: Phase 3 re-triggered with adjusted queries [threshold: <= 2 retries]
+- [ ] On FAIL after max retries: partial research returned with explicit warning [threshold: >= 1 warning]
+- [ ] Quality score written to research document's metadata section [threshold: >= 1]
+- [ ] Validation result is structured YAML, not free text [threshold: >= 1 valid YAML]
+
+## Veto Conditions
+
+| ID | Condition | Check | Result |
+|----|-----------|-------|--------|
+| VETO-DRV-001 | Research quality score must meet minimum threshold | Validate quality_score >= 60% | VETO - BLOCK. Retry research with adjusted queries (max 2 retries) or flag for manual enrichment. |
+
+## Related Documents
+
+| Document | Relationship |
+|----------|-------------|
+| `deep-research-pre-agent.md` | Parent stub (orchestrator) |
+| `deep-research-check-local-knowledge.md` | Phase 1 |
+| `deep-research-generate-prompt.md` | Phase 2 |
+| `deep-research-execute.md` | Phase 3 (produces the document validated here) |
+| `checklists/deep-research-quality.md` | Quality checklist |

package/pro/squads/squad-creator-pro/tasks/delete-squad.md

@@ -0,0 +1,104 @@
+<!-- SINKRA_TASK_METADATA:START -->
+```yaml
+sinkra_task_metadata:
+  task_id: delete-squad
+  task_name: Delete Squad Artifact Surface
+  status: pending
+  responsible_executor: '@squad-chief'
+  execution_type: Hybrid
+  estimated_time: 30m
+  domain: Tactical
+  input:
+  - Consultar a seção de inputs no corpo da task
+  output:
+  - Consultar a seção de outputs no corpo da task
+  action_items:
+  - Executar os passos documentados no corpo da task
+  acceptance_criteria:
+  - Backup criado e validado
+  - Escopo da remoção confirmado explicitamente
+  - Journey log e relatório de arquivamento gerados
+  output_persistence: transient_output
+  accountable_id: Human:Squad_Operator
+  accountability_scope: full
+  escalation_priority: medium
+```
+<!-- SINKRA_TASK_METADATA:END -->
+
+<!-- SINKRA_CONTRACT:START -->
+```yaml
+sinkra_contract:
+  Domain: Operational
+  atomic_layer: Atom
+  executor: Agent
+  pre_condition: "inputs, dependências e artefatos prévios resolvidos antes de iniciar a execução."
+  post_condition: "output principal gerado, validado e pronto para handoff da próxima fase."
+  performance: "executar dentro do SLA declarado, registrar erro explicitamente e escalar via handoff sem falha silenciosa."
+```
+<!-- SINKRA_CONTRACT:END -->
+
+
+# Task: Delete Squad Artifact Surface
+
+## Task Anatomy
+
+| Field | Value |
+|-------|-------|
+| **Task ID** | `delete-squad` |
+| **Version** | `1.0.0` |
+| **Status** | `active` |
+| **Responsible Executor** | `@squad-chief` |
+| **Execution Type** | `Hybrid` |
+
+## Metadata
+
+```yaml
+id: delete-squad
+name: "Delete Squad Artifact Surface"
+category: maintenance
+agent: squad-chief
+elicit: true
+autonomous: false
+description: "Executa a retirada controlada de um squad ou artefato do pack, com backup, trilha auditável e atualização do journey log."
+task_name: "Delete Squad Artifact Surface"
+status: "active"
+responsible_executor: "@squad-chief"
+execution_type: "Hybrid"
+estimated_time: "30-60m"
+domain: Tactical
+input:
+  - "squad_name"
+  - "delete_scope"
+  - "backup_path"
+output:
+  - "archive_report"
+  - "journey_log_entry"
+action_items:
+  - "Criar backup verificável da superfície afetada"
+  - "Remover ou arquivar apenas os artefatos aprovados"
+  - "Atualizar contratos, registries e journey log"
+acceptance_criteria:
+  - "Backup criado antes de qualquer remoção"
+  - "Journey log atualizado com motivo e sucessor quando aplicável"
+  - "Nenhum artefato ativo órfão permanece após a remoção"
+output_persistence: "transient_output"
+accountable_id: "Human:Squad_Operator"
+accountability_scope: "full"
+escalation_priority: "high"
+```
+
+## Purpose
+
+Oferecer um delete contextual para lifecycle do pack sem usar remoção cega. A task existe para cumprir o contrato de lifecycle do validator e para garantir que qualquer retirada seja reversível, auditável e deliberada.
+
+## Command Contract
+
+```text
+*delete-squad {squad_name} [--scope artifacts|runtime|full] [--backup-path {path}]
+```
+
+## Acceptance Criteria
+
+- [ ] Backup criado e validado
+- [ ] Escopo da remoção confirmado explicitamente
+- [ ] Journey log e relatório de arquivamento gerados

package/pro/squads/squad-creator-pro/tasks/discover-tools-execute.md

@@ -0,0 +1,293 @@
+<!-- SINKRA_TASK_METADATA:START -->
+```yaml
+sinkra_task_metadata:
+  task_id: discover-tools-execute
+  task_name: Discover Tools Execute (Consolidated Pipeline Proxy)
+  status: pending
+  responsible_executor: '@squad-chief'
+  execution_type: Agent
+  estimated_time: 45m
+  domain: Operational
+  input:
+  - domain (string, required)
+  - use_cases (list, required)
+  - existing_tools (list, optional)
+  - capability_gaps (list, optional)
+  - budget_tier (enum, optional)
+  - priority_capabilities (list, optional)
+  - search_depth (enum, optional)
+  - scope_hint (string, optional)
+  output:
+  - '.aiox/squad-runtime/discovery/{domain}/scan-results.yaml'
+  - '.aiox/squad-runtime/discovery/{domain}/classified-tools.yaml'
+  - '.aiox/squad-runtime/discovery/{domain}/evaluated-tools.yaml'
+  - '.aiox/squad-runtime/discovery/{domain}/recommendations.yaml'
+  - '.aiox/squad-runtime/discovery/{domain}/tool-discovery-report.md'
+  - '.aiox/squad-runtime/discovery/{domain}/capability-tools.yaml'
+  - '.aiox/squad-runtime/discovery/{domain}/tool-integration-plan.md'
+  action_items:
+  - 'Step 1: Scan -- delegate to squad-creator base scan task'
+  - 'Step 2: Classify -- delegate to squad-creator base classify task'
+  - 'Step 3: Evaluate -- delegate to squad-creator base evaluate task'
+  - 'Step 4: Recommend -- delegate to squad-creator base recommend task'
+  - 'Step 5: Report -- delegate to squad-creator base report task'
+  acceptance_criteria:
+  - O pro não mantém uma segunda implementação de nenhuma fase do pipeline
+  - Cada step delega ao owner canônico no squad-creator base
+  - Os artefatos locais do pro (tool-registry, internal-infrastructure-library) entram como overlay
+  - Todos os outputs intermediários e finais são produzidos nos paths esperados
+  - O scoring é herdado do base (nenhuma lógica paralela de score)
+  - A recomendação é determinada pelo owner canônico do base
+  - A geração dos artefatos finais continua centralizada no base
+  - O workflow do pro possui owner atômico explícito em cada fase
+  output_persistence: transient_output
+  accountable_id: Human:Squad_Operator
+  accountability_scope: review_only
+  escalation_priority: medium
+  coherence_threshold: 0.95
+  error_behavior: raise
+```
+<!-- SINKRA_TASK_METADATA:END -->
+
+<!-- SINKRA_CONTRACT:START -->
+```yaml
+sinkra_contract:
+  Domain: Tactical
+  atomic_layer: Atom
+  executor: Agent
+  pre_condition: "inputs, dependências e artefatos prévios resolvidos antes de iniciar a execução."
+  post_condition: "output principal gerado, validado e pronto para handoff da próxima fase."
+  performance: "executar dentro do SLA declarado, registrar erro explicitamente e escalar via handoff sem falha silenciosa."
+```
+<!-- SINKRA_CONTRACT:END -->
+
+
+# Task: Discover Tools Execute (Consolidated Pipeline Proxy)
+
+**Task ID:** discover-tools-execute
+**Version:** 1.0.0
+**Purpose:** Execute the complete tool discovery pipeline (scan, classify, evaluate, recommend, report) by delegating each phase to the canonical owner in `squad-creator` base
+**Executor:** Agent (@squad-chief)
+**Mode:** Delegation-first
+**Quality Standard:** Herdado das tasks base em `squads/squad-creator/tasks/`
+
+---
+
+## Task Anatomy
+
+| Field | Value |
+|-------|-------|
+| **Executor** | Agent (@squad-chief) |
+| **Inputs** | domain, use_cases, existing_tools, capability_gaps, budget_tier, priority_capabilities, search_depth, scope_hint |
+| **Outputs** | scan-results, classified-tools, evaluated-tools, recommendations, tool-discovery-report, capability-tools, tool-integration-plan |
+| **Completion Criteria** | All 5 pipeline steps delegated and all outputs produced in `.aiox/squad-runtime/discovery/{domain}/` |
+| **Guardrails** | No parallel logic -- all execution delegated to squad-creator base tasks |
+
+---
+
+## Purpose
+
+This consolidated proxy replaces the 5 individual wrapper stubs (discover-tools-scan, discover-tools-classify, discover-tools-evaluate, discover-tools-recommend, discover-tools-report) that existed as pure delegation wrappers.
+
+Each step delegates to the canonical owner in `squad-creator` base. The `squad-creator-pro` adds only:
+
+1. Pro overlay context (tool-registry, internal-infrastructure-library)
+2. Pipeline sequencing within a single task
+3. Output reconciliation
+
+**Non-negotiable rule:** The `squad-creator-pro` MUST NOT maintain parallel implementations of any pipeline phase.
+
+---
+
+## Inputs
+
+| Parameter | Type | Required | Description |
+|-----------|------|----------|-------------|
+| `domain` | string | Yes | Dominio ou squad alvo |
+| `use_cases` | list | Yes | Casos de uso principais |
+| `existing_tools` | list | No | Ferramentas ja conhecidas |
+| `capability_gaps` | list | No | Gaps ja percebidos |
+| `budget_tier` | enum | No | `free_only`, `low_cost`, `enterprise` (default: `low_cost`) |
+| `priority_capabilities` | list | No | Capacidades a priorizar |
+| `search_depth` | enum | No | `quick`, `standard`, `exhaustive` (default: `standard`) |
+| `scope_hint` | string | No | Clarificacao para dominios ambiguos |
+
+---
+
+## Preconditions
+
+- [ ] `squads/squad-creator/tasks/discover-tools-scan.md` exists
+- [ ] `squads/squad-creator/tasks/discover-tools-classify.md` exists
+- [ ] `squads/squad-creator/tasks/discover-tools-evaluate.md` exists
+- [ ] `squads/squad-creator/tasks/discover-tools-recommend.md` exists
+- [ ] `squads/squad-creator/tasks/discover-tools-report.md` exists
+- [ ] `data/internal-infrastructure-library.yaml` exists in squad-creator-pro
+- [ ] `data/tool-registry.yaml` exists in squad-creator-pro
+
+---
+
+## Workflow
+
+### Step 1: Scan -- Structural Discovery
+
+Delegate structural scanning of tools to the base canonical task.
+
+```yaml
+delegate_to_base:
+  task: "squads/squad-creator/tasks/discover-tools-scan.md"
+  payload:
+    - domain
+    - use_cases
+    - existing_tools
+    - capability_gaps
+    - scope_hint
+  overlay_inputs:
+    - internal_infrastructure_library: "data/internal-infrastructure-library.yaml"
+    - tool_registry: "data/tool-registry.yaml"
+  prohibition:
+    - "Do NOT reimplement scan logic inside squad-creator-pro"
+  output: ".aiox/squad-runtime/discovery/{domain}/scan-results.yaml"
+```
+
+### Step 2: Classify -- Candidate Classification
+
+Delegate classification of discovered tool candidates to the base canonical task.
+
+```yaml
+delegate_to_base:
+  task: "squads/squad-creator/tasks/discover-tools-classify.md"
+  payload:
+    - scan_output: "{step_1_output}"
+    - domain
+    - use_cases
+    - budget_tier
+  overlay_rules:
+    - "Honor external-only gaps produced by the pro scan step"
+    - "Keep pro-local exclusions and internal-first decisions"
+  prohibition:
+    - "Do NOT fork classification logic in squad-creator-pro"
+  output: ".aiox/squad-runtime/discovery/{domain}/classified-tools.yaml"
+```
+
+### Step 3: Evaluate -- Scoring and Compatibility
+
+Delegate scoring and compatibility evaluation to the base canonical task.
+
+```yaml
+delegate_to_base:
+  task: "squads/squad-creator/tasks/discover-tools-evaluate.md"
+  payload:
+    - classified_tools: "{step_2_output}"
+    - prioritized_gaps: "{from_step_1}"
+    - budget_tier
+  prohibition:
+    - "Do NOT fork tool scoring in squad-creator-pro"
+  output: ".aiox/squad-runtime/discovery/{domain}/evaluated-tools.yaml"
+```
+
+### Step 4: Recommend -- Prioritization and Decision
+
+Delegate recommendation and prioritization to the base canonical task.
+
+```yaml
+delegate_to_base:
+  task: "squads/squad-creator/tasks/discover-tools-recommend.md"
+  payload:
+    - evaluated_tools: "{step_3_output}"
+    - scan_output: "{step_1_output}"
+    - domain
+  prohibition:
+    - "Do NOT keep a second impact-vs-effort matrix in squad-creator-pro"
+  output: ".aiox/squad-runtime/discovery/{domain}/recommendations.yaml"
+```
+
+### Step 5: Report -- Final Artifact Generation
+
+Delegate final report and artifact generation to the base canonical task.
+
+```yaml
+delegate_to_base:
+  task: "squads/squad-creator/tasks/discover-tools-report.md"
+  payload:
+    - recommend_output: "{step_4_output}"
+    - scan_output: "{step_1_output}"
+    - domain
+  prohibition:
+    - "Do NOT duplicate report generation logic in squad-creator-pro"
+  outputs:
+    - ".aiox/squad-runtime/discovery/{domain}/tool-discovery-report.md"
+    - ".aiox/squad-runtime/discovery/{domain}/capability-tools.yaml"
+    - ".aiox/squad-runtime/discovery/{domain}/tool-integration-plan.md"
+```
+
+---
+
+## Output
+
+```yaml
+output:
+  name: discover_tools_pipeline_result
+  format: yaml
+  structure:
+    pipeline_steps:
+      scan:
+        delegated_task: "squads/squad-creator/tasks/discover-tools-scan.md"
+        output: ".aiox/squad-runtime/discovery/{domain}/scan-results.yaml"
+      classify:
+        delegated_task: "squads/squad-creator/tasks/discover-tools-classify.md"
+        output: ".aiox/squad-runtime/discovery/{domain}/classified-tools.yaml"
+      evaluate:
+        delegated_task: "squads/squad-creator/tasks/discover-tools-evaluate.md"
+        output: ".aiox/squad-runtime/discovery/{domain}/evaluated-tools.yaml"
+      recommend:
+        delegated_task: "squads/squad-creator/tasks/discover-tools-recommend.md"
+        output: ".aiox/squad-runtime/discovery/{domain}/recommendations.yaml"
+      report:
+        delegated_task: "squads/squad-creator/tasks/discover-tools-report.md"
+        outputs:
+          - ".aiox/squad-runtime/discovery/{domain}/tool-discovery-report.md"
+          - ".aiox/squad-runtime/discovery/{domain}/capability-tools.yaml"
+          - ".aiox/squad-runtime/discovery/{domain}/tool-integration-plan.md"
+    domain: "{domain}"
+    execution_mode: "consolidated-proxy"
+    status: "delegated"
+```
+
+---
+
+## Acceptance Criteria
+
+- [ ] O pro nao mantem uma segunda implementacao de nenhuma fase do pipeline
+- [ ] Cada step delega ao owner canonico no squad-creator base
+- [ ] Os artefatos locais do pro (tool-registry, internal-infrastructure-library) entram como overlay
+- [ ] Todos os outputs intermediarios e finais sao produzidos nos paths esperados
+- [ ] O scoring e herdado do base (nenhuma logica paralela de score)
+- [ ] A recomendacao e determinada pelo owner canonico do base
+- [ ] A geracao dos artefatos finais continua centralizada no base
+- [ ] O workflow do pro possui owner atomico explicito em cada fase
+
+---
+
+## Veto Conditions
+
+- Any base task in `squads/squad-creator/tasks/discover-tools-*.md` does not exist
+- `domain` not provided
+- `use_cases` not provided
+
+---
+
+## Related Documents
+
+- `squads/squad-creator/tasks/discover-tools-scan.md` -- canonical owner (scan)
+- `squads/squad-creator/tasks/discover-tools-classify.md` -- canonical owner (classify)
+- `squads/squad-creator/tasks/discover-tools-evaluate.md` -- canonical owner (evaluate)
+- `squads/squad-creator/tasks/discover-tools-recommend.md` -- canonical owner (recommend)
+- `squads/squad-creator/tasks/discover-tools-report.md` -- canonical owner (report)
+- `workflows/wf-discover-tools.yaml` -- workflow consumer in pro
+- `data/tool-registry.yaml` -- tool catalog and memory
+- `data/internal-infrastructure-library.yaml` -- pro infrastructure hints
+
+---
+
+_Task Version: 1.0.0_
+_Role: consolidated pipeline proxy replacing 5 individual wrapper stubs_