recomendar-estrategias

name	recomendar-estrategias
description	Maps gaps per capability to the 7 strategies S1-S7 and generates prioritized action plan with specific technologies. Reads saida/gaps.json + framework.json. Generates saida/recomendacoes.json. Use when user asks to "recomendar estratégias", "mapear ações", "que iniciativas devo priorizar".

Skill: Recommend strategies S1–S7

When to use

• After /gap-analysis (depends on saida/gaps.json).
• When client wants an actionable plan, not just a diagnosis.

Inputs

• saida/gaps.json — prioritized gaps
• framework.json::strategies — definition of the 7 strategies
• framework.json::technologies_per_strategy — technologies per strategy
• framework.json::pillars[].capabilities[].strategies — cap→strategies mapping

Expected output

• saida/recomendacoes.json — grouped by strategy + ranking
• Brief chat message (PT-BR): top 3 strategies by cumulative impact.

Algorithm

1. Aggregate gaps per strategy

For each gap in gaps.json::gaps:

for each strategy_id in gap.strategies:
    estrategias[strategy_id].related_gaps.append(gap)
    estrategias[strategy_id].cumulative_priority += gap.priority_score
    estrategias[strategy_id].max_priority = max(P0, P1, P2, P3 of gaps)

2. Rank strategies

Sort by cumulative_priority desc. Tie → higher max_priority → more related_gaps.

3. Technologies and horizon

For each strategy:

• Pull technologies_per_strategy[strategy_id] from framework.json
• Determine aggregated horizon from the most critical gap in the group (P0=30d, P1=quarter, P2=semester)

4. Generate concrete actions (deterministic templates — DO NOT invent)

Use these templates based on the strategy (output text in PT-BR):

Strategy	Recommended initial action
S1 GitHub Migration	"Inventário de repositórios atuais → plano de migração para GitHub Enterprise Cloud em 3 ondas."
S2 Foundry + SRE	"Definir SLOs/SLIs para serviços críticos; implantar Azure Monitor + dashboards Grafana."
S3 App Modernization	"Selecionar 1–2 apps piloto; replatforming para Azure Container Apps + IaC com Terraform."
S4 AI Applications	"Identificar 2 casos de uso de alto ROI; PoC com Azure OpenAI + Prompt Flow."
S5 Copilot Acceleration	"Rollout Copilot Enterprise em 2 squads piloto; medir adoção e produtividade DORA por 8 semanas."
S6 Agentic Activation	"Pilot Semantic Kernel para 1 workflow interno; estabelecer guardrails e observabilidade."
S7 Security & Governance	"Habilitar GitHub Advanced Security em todos repos; gerar SBOM dos serviços críticos."

saida/recomendacoes.json schema

{
  "metadata": {
    "computed_at": "2026-05-08T14:30:00Z",
    "based_on": "saida/gaps.json"
  },
  "ranked_strategies": [
    {
      "rank": 1,
      "strategy_id": "S5",
      "strategy_name": "GitHub Copilot Acceleration",
      "cumulative_priority": 8.42,
      "max_priority": "P0 — Crítico",
      "horizon": "30 dias",
      "related_capabilities": [
        {"id": "P1-C1", "name_pt_br": "Assistentes de Codificação IA", "gap_size": 1.4, "priority": "P1 — Alto"},
        {"id": "P1-C8", "name_pt_br": "Medição de Produtividade", "gap_size": 1.5, "priority": "P1 — Alto"}
      ],
      "technologies": [
        {"name": "GitHub Copilot Enterprise", "purpose": "AI-assisted coding across development teams"}
      ],
      "first_action": "Rollout Copilot Enterprise em 2 squads piloto; medir adoção e produtividade DORA por 8 semanas.",
      "expected_outcome": "Subir capabilities P1-C1 e P1-C8 para L3 em ~2 trimestres."
    }
  ],
  "skipped_strategies": ["S1"]
}

skipped_strategies = strategies that didn't appear in any gap (client already mature in them or related capabilities not answered).

Report in chat (PT-BR)

✓ Recomendações → saida/recomendacoes.json

Top 3 estratégias por impacto cumulativo:
  🥇 S5 — GitHub Copilot Acceleration (priority 8.42, 4 capabilities)
     → Rollout Copilot em 2 squads piloto, medir DORA por 8 sem.
  🥈 S7 — Security & Governance (priority 5.10, 2 capabilities)
     → Habilitar GHAS em todos repos + SBOM dos serviços críticos.
  🥉 S6 — Agentic Activation (priority 2.94, 1 capability)
     → Pilot Semantic Kernel para 1 workflow interno.

Não recomendadas (sem gaps relevantes): S1
Próximo: /gerar-relatorio

Constraints

• DO NOT INVENT actions outside the table above — if in doubt, write "Detalhar com arquiteto Microsoft GBB".
• Strategies with cumulative_priority < 0.9 (only P3 gaps) → move to skipped_strategies with note "monitorar".
• Technologies must come EXCLUSIVELY from framework.json::technologies_per_strategy.
• Keep PT-BR in human-facing fields; technical IDs and names can stay in English.