@luanpdd/kit-mcp 1.21.0 → 1.26.0

package/kit/commands/sync-main.md

@@ -1,68 +1,68 @@
----
-name: sync-main
-description: Atualiza a branch local com os commits da main. Se houver conflitos, pergunta qual priorizar.
-allowed-tools:
-  - Bash
----
-
-<safety_rules>
-NUNCA execute git push para main ou origin/main.
-NUNCA mude para a branch main (git checkout main / git switch main).
-Este comando só traz código da main para a branch local — nunca o contrário.
-</safety_rules>
-
-<process>
-
-## 1 — Proteção: bloquear se estiver na main
-
-```bash
-git branch --show-current
-```
-
-Se a branch atual for `main`, encerre com:
-> "⛔ Você está na branch main. Mude para uma branch de feature antes de continuar."
-
-## 2 — Buscar atualizações e verificar se há algo novo
-
-```bash
-git fetch origin
-git log --oneline origin/main ^HEAD
-```
-
-Se não houver nenhum commit novo em `origin/main`, informe:
-> "Sua branch já está atualizada com a main."
-E encerre.
-
-## 3 — Aplicar os commits da main na branch local
-
-```bash
-git merge origin/main
-```
-
-### Se não houver conflitos:
-Mostre um resumo com:
-- Quantos commits foram integrados
-- Lista dos arquivos modificados (`git diff --name-only HEAD~1 HEAD` ou similar)
-
-Encerre.
-
-### Se houver conflitos:
-
-Liste os arquivos conflitantes e pergunte:
-> "Encontrei conflitos nos arquivos abaixo. Qual é a prioridade?
-> [1] main — usa o código da main nos conflitos
-> [2] local — mantém o seu código nos conflitos"
-
-Aguarde a resposta.
-
-**Se [1] main:**
-Para cada arquivo conflitante: `git checkout --theirs <arquivo>`
-Depois: `git add . && git commit`
-
-**Se [2] local:**
-Para cada arquivo conflitante: `git checkout --ours <arquivo>`
-Depois: `git add . && git commit`
-
-Mostre o resumo final: branch atual, status limpo, commits integrados.
-
-</process>
+---
+name: sync-main
+description: Atualiza a branch local com os commits da main. Se houver conflitos, pergunta qual priorizar.
+allowed-tools:
+  - Bash
+---
+
+<safety_rules>
+NUNCA execute git push para main ou origin/main.
+NUNCA mude para a branch main (git checkout main / git switch main).
+Este comando só traz código da main para a branch local — nunca o contrário.
+</safety_rules>
+
+<process>
+
+## 1 — Proteção: bloquear se estiver na main
+
+```bash
+git branch --show-current
+```
+
+Se a branch atual for `main`, encerre com:
+> "⛔ Você está na branch main. Mude para uma branch de feature antes de continuar."
+
+## 2 — Buscar atualizações e verificar se há algo novo
+
+```bash
+git fetch origin
+git log --oneline origin/main ^HEAD
+```
+
+Se não houver nenhum commit novo em `origin/main`, informe:
+> "Sua branch já está atualizada com a main."
+E encerre.
+
+## 3 — Aplicar os commits da main na branch local
+
+```bash
+git merge origin/main
+```
+
+### Se não houver conflitos:
+Mostre um resumo com:
+- Quantos commits foram integrados
+- Lista dos arquivos modificados (`git diff --name-only HEAD~1 HEAD` ou similar)
+
+Encerre.
+
+### Se houver conflitos:
+
+Liste os arquivos conflitantes e pergunte:
+> "Encontrei conflitos nos arquivos abaixo. Qual é a prioridade?
+> [1] main — usa o código da main nos conflitos
+> [2] local — mantém o seu código nos conflitos"
+
+Aguarde a resposta.
+
+**Se [1] main:**
+Para cada arquivo conflitante: `git checkout --theirs <arquivo>`
+Depois: `git add . && git commit`
+
+**Se [2] local:**
+Para cada arquivo conflitante: `git checkout --ours <arquivo>`
+Depois: `git add . && git commit`
+
+Mostre o resumo final: branch atual, status limpo, commits integrados.
+
+</process>

package/kit/commands/validar-fase.md

@@ -1,35 +1,35 @@
----
-name: validar-fase
-description: Audita retroativamente e preenche lacunas de validação Nyquist para uma fase concluída
-argument-hint: "[número da fase]"
-allowed-tools:
-  - Read
-  - Write
-  - Edit
-  - Bash
-  - Glob
-  - Grep
-  - Task
-  - AskUserQuestion
----
-<objective>
-Audita a cobertura de validação Nyquist para uma fase concluída. Três estados:
-- (A) VALIDATION.md existe — auditar e preencher lacunas
-- (B) Sem VALIDATION.md, SUMMARY.md existe — reconstruir a partir de artefatos
-- (C) Fase não executada — sair com orientação
-
-Saída: VALIDATION.md atualizado + arquivos de teste gerados.
-</objective>
-
-<execution_context>
-@./.claude/framework/workflows/validate-phase.md
-</execution_context>
-
-<context>
-Fase: $ARGUMENTS — opcional, padrão é a última fase concluída.
-</context>
-
-<process>
-Execute @./.claude/framework/workflows/validate-phase.md.
-Preserve todos os gates do workflow.
-</process>
+---
+name: validar-fase
+description: Audita retroativamente e preenche lacunas de validação Nyquist para uma fase concluída
+argument-hint: "[número da fase]"
+allowed-tools:
+  - Read
+  - Write
+  - Edit
+  - Bash
+  - Glob
+  - Grep
+  - Task
+  - AskUserQuestion
+---
+<objective>
+Audita a cobertura de validação Nyquist para uma fase concluída. Três estados:
+- (A) VALIDATION.md existe — auditar e preencher lacunas
+- (B) Sem VALIDATION.md, SUMMARY.md existe — reconstruir a partir de artefatos
+- (C) Fase não executada — sair com orientação
+
+Saída: VALIDATION.md atualizado + arquivos de teste gerados.
+</objective>
+
+<execution_context>
+@./.claude/framework/workflows/validate-phase.md
+</execution_context>
+
+<context>
+Fase: $ARGUMENTS — opcional, padrão é a última fase concluída.
+</context>
+
+<process>
+Execute @./.claude/framework/workflows/validate-phase.md.
+Preserve todos os gates do workflow.
+</process>

package/kit/commands/verificar-tarefas.md

@@ -1,45 +1,45 @@
----
-name: verificar-tarefas
-description: Lista todos os todos pendentes e seleciona um para trabalhar
-argument-hint: [filtro de área]
-allowed-tools:
-  - Read
-  - Write
-  - Bash
-  - AskUserQuestion
----
-
-<objective>
-Lista todos os todos pendentes, permite seleção, carrega o contexto completo do todo selecionado e encaminha para a ação adequada.
-
-Encaminha para o workflow check-todos que cuida de:
-- Contagem e listagem de todos com filtro por área
-- Seleção interativa com carregamento completo de contexto
-- Checagem de correlação com roadmap
-- Roteamento de ação (trabalhar agora, adicionar à fase, brainstorm, criar fase)
-- Atualizações do STATE.md e commits git
-</objective>
-
-<execution_context>
-@./.claude/framework/workflows/check-todos.md
-</execution_context>
-
-<context>
-Argumentos: $ARGUMENTS (filtro de área opcional)
-
-Estado dos todos e correlação com roadmap são carregados no workflow usando `init todos` e leituras específicas.
-</context>
-
-<process>
-**Seguir o workflow check-todos** de `@./.claude/framework/workflows/check-todos.md`.
-
-O workflow cuida de toda a lógica, incluindo:
-1. Verificação de existência de todos
-2. Filtro por área
-3. Listagem e seleção interativa
-4. Carregamento completo de contexto com resumos de arquivo
-5. Checagem de correlação com roadmap
-6. Oferta e execução de ação
-7. Atualizações do STATE.md
-8. Commits git
+---
+name: verificar-tarefas
+description: Lista todos os todos pendentes e seleciona um para trabalhar
+argument-hint: [filtro de área]
+allowed-tools:
+  - Read
+  - Write
+  - Bash
+  - AskUserQuestion
+---
+
+<objective>
+Lista todos os todos pendentes, permite seleção, carrega o contexto completo do todo selecionado e encaminha para a ação adequada.
+
+Encaminha para o workflow check-todos que cuida de:
+- Contagem e listagem de todos com filtro por área
+- Seleção interativa com carregamento completo de contexto
+- Checagem de correlação com roadmap
+- Roteamento de ação (trabalhar agora, adicionar à fase, brainstorm, criar fase)
+- Atualizações do STATE.md e commits git
+</objective>
+
+<execution_context>
+@./.claude/framework/workflows/check-todos.md
+</execution_context>
+
+<context>
+Argumentos: $ARGUMENTS (filtro de área opcional)
+
+Estado dos todos e correlação com roadmap são carregados no workflow usando `init todos` e leituras específicas.
+</context>
+
+<process>
+**Seguir o workflow check-todos** de `@./.claude/framework/workflows/check-todos.md`.
+
+O workflow cuida de toda a lógica, incluindo:
+1. Verificação de existência de todos
+2. Filtro por área
+3. Listagem e seleção interativa
+4. Carregamento completo de contexto com resumos de arquivo
+5. Checagem de correlação com roadmap
+6. Oferta e execução de ação
+7. Atualizações do STATE.md
+8. Commits git
 </process>

package/kit/commands/verificar-trabalho.md

@@ -1,64 +1,64 @@
----
-name: verificar-trabalho
-description: Valida funcionalidades construídas através de UAT conversacional
-argument-hint: "[número da fase, ex: '4']"
-allowed-tools:
-  - Read
-  - Bash
-  - Glob
-  - Grep
-  - Edit
-  - Write
-  - Task
----
-<objective>
-Valida funcionalidades construídas através de testes conversacionais com estado persistente.
-
-Propósito: Confirmar que o que o Claude construiu realmente funciona da perspectiva do usuário. Um teste por vez, respostas em texto simples, sem interrogatório. Quando problemas são encontrados, automaticamente diagnostica, planeja correções e prepara para execução.
-
-Saída: {phase_num}-UAT.md rastreando todos os resultados de teste. Se problemas encontrados: lacunas diagnosticadas, planos de correção verificados prontos para /executar-fase
-</objective>
-
-<execution_context>
-@./.claude/framework/workflows/verify-work.md
-@./.claude/framework/templates/UAT.md
-</execution_context>
-
-<context>
-Fase: $ARGUMENTS (opcional)
-- Se fornecido: Testar fase específica (ex: "4")
-- Se não fornecido: Verificar sessões ativas ou solicitar fase
-
-Arquivos de contexto são resolvidos dentro do workflow (`init verify-work`) e delegados via blocos `<files_to_read>`.
-</context>
-
-<process>
-Execute o workflow verify-work de @./.claude/framework/workflows/verify-work.md do início ao fim.
-Preserve todos os gates do workflow (gerenciamento de sessão, apresentação de testes, diagnóstico, planejamento de correções, roteamento).
-</process>
-
-<observability_integration>
-**Integração com Core Analysis Loop em logs reais (v1.9):**
-
-Quando `workflow.observability_uat_validation = true` (default) e o projeto tem MCP Supabase disponível, o workflow inclui passo de validação observacional após UAT conversacional:
-
-1. Para cada test passing no UAT, invocar o agente [`incident-investigator`](../agents/incident-investigator.md) em modo "validation":
-   ```
-   Task(subagent_type="incident-investigator", prompt="
-     mode: validation
-     symptom: validar que feature de Phase {N} efetivamente emitiu spans/eventos esperados
-     time_window: última 1h
-     expected_attributes: {result.success: true, build_id: {current_build}}
-   ")
-   ```
-2. O agente queryará via `mcp__supabase__get_logs` ou `mcp__supabase__execute_sql` para confirmar:
-   - Spans com nome esperado existem
-   - Atributos canônicos foram emitidos (não só código existe — comportamento real)
-   - `result.success` baseline está dentro do esperado
-3. Se confirmado: UAT.md inclui evidência observacional (não só "funciona em UI"). Status `passed` ✓.
-4. Se ausente: UAT.md flag `human_needed` com sugestão "verificar instrumentação não está pegando".
-
-**Skill consultada:** [`core-analysis-loop`](../skills/core-analysis-loop/SKILL.md) — para o caso de UAT falhar e precisar investigar.
-
-**REQ:** INT-FW-03.
-</observability_integration>
+---
+name: verificar-trabalho
+description: Valida funcionalidades construídas através de UAT conversacional
+argument-hint: "[número da fase, ex: '4']"
+allowed-tools:
+  - Read
+  - Bash
+  - Glob
+  - Grep
+  - Edit
+  - Write
+  - Task
+---
+<objective>
+Valida funcionalidades construídas através de testes conversacionais com estado persistente.
+
+Propósito: Confirmar que o que o Claude construiu realmente funciona da perspectiva do usuário. Um teste por vez, respostas em texto simples, sem interrogatório. Quando problemas são encontrados, automaticamente diagnostica, planeja correções e prepara para execução.
+
+Saída: {phase_num}-UAT.md rastreando todos os resultados de teste. Se problemas encontrados: lacunas diagnosticadas, planos de correção verificados prontos para /executar-fase
+</objective>
+
+<execution_context>
+@./.claude/framework/workflows/verify-work.md
+@./.claude/framework/templates/UAT.md
+</execution_context>
+
+<context>
+Fase: $ARGUMENTS (opcional)
+- Se fornecido: Testar fase específica (ex: "4")
+- Se não fornecido: Verificar sessões ativas ou solicitar fase
+
+Arquivos de contexto são resolvidos dentro do workflow (`init verify-work`) e delegados via blocos `<files_to_read>`.
+</context>
+
+<process>
+Execute o workflow verify-work de @./.claude/framework/workflows/verify-work.md do início ao fim.
+Preserve todos os gates do workflow (gerenciamento de sessão, apresentação de testes, diagnóstico, planejamento de correções, roteamento).
+</process>
+
+<observability_integration>
+**Integração com Core Analysis Loop em logs reais (v1.9):**
+
+Quando `workflow.observability_uat_validation = true` (default) e o projeto tem MCP Supabase disponível, o workflow inclui passo de validação observacional após UAT conversacional:
+
+1. Para cada test passing no UAT, invocar o agente [`incident-investigator`](../agents/incident-investigator.md) em modo "validation":
+   ```
+   Task(subagent_type="incident-investigator", prompt="
+     mode: validation
+     symptom: validar que feature de Phase {N} efetivamente emitiu spans/eventos esperados
+     time_window: última 1h
+     expected_attributes: {result.success: true, build_id: {current_build}}
+   ")
+   ```
+2. O agente queryará via `mcp__supabase__get_logs` ou `mcp__supabase__execute_sql` para confirmar:
+   - Spans com nome esperado existem
+   - Atributos canônicos foram emitidos (não só código existe — comportamento real)
+   - `result.success` baseline está dentro do esperado
+3. Se confirmado: UAT.md inclui evidência observacional (não só "funciona em UI"). Status `passed` ✓.
+4. Se ausente: UAT.md flag `human_needed` com sugestão "verificar instrumentação não está pegando".
+
+**Skill consultada:** [`core-analysis-loop`](../skills/core-analysis-loop/SKILL.md) — para o caso de UAT falhar e precisar investigar.
+
+**REQ:** INT-FW-03.
+</observability_integration>

package/kit/file-manifest.json

@@ -1,41 +1,43 @@
 {
-  "version": "1.21.0",
-  "timestamp": "2026-05-10T15:32:12.077Z",
+  "version": "1.26.0",
+  "timestamp": "2026-05-11T04:13:53.239Z",
   "files": {
     "COMANDOS.md": "d24ec61a6ec35db314cc5f2ae287bfb927b794789c8f1d558c55862f5e6534b2",
     "COMPATIBILITY.md": "794e336a87045cdf0161785b9a7a0975a49abbd80bdd816b8852251fcc8126ca",
-    "README.md": "358e735e8b9bd044fe7c6442014f942a0c9de182306448a06f47342ae57e6582",
+    "README.md": "a2adf08b15a433516534fb8aa3569fa2c151e48cdafcf9d1b13a26e46b95b357",
     "agents/advisor-researcher.md": "2ecc52247af1f379c7d0a85847a966095f71d8e5a6140f62611f04c9e4fe264b",
     "agents/ai-mutation-tester.md": "bbdedc2d41440340c403bf4b71fb902eec057f3771667703c32309885aa4f4db",
     "agents/assumptions-analyzer.md": "b730268a8c588bc16e30330df874f64d1dc0bbd4573e696f7a316ea9e4c1f57c",
-    "agents/audit-log-implementer.md": "b8520400c6dfd27ec3675ee84fdac6f8bddb314e4800b6edcac9bbc761a726ae",
+    "agents/audit-log-implementer.md": "9ef8239b0186b3e3526f6c61a8fa2cca73738bca1cdfd0d85e4187efda6156d1",
+    "agents/auditor-consistencia-isolamento.md": "0d067455d39a8ad3fcc4eccea6fecaeaf3ba9d724b29e58c3612d33dfe6b6236",
     "agents/b2b-saas-architect.md": "6e613518b6c101a1d7cf0d86ef9337347662db863e47a492c3a2415a9b43ec3d",
     "agents/burn-rate-forecaster.md": "a4f2efccc7073ef0ab1a225895f68dff7c95d4f9e1c04319569cad47c13dffac",
     "agents/cascading-failures-auditor.md": "6c44929a90a2fc59d3ba59dd67814a85dc08f1b86d41933c5f19cefc8d33bdd5",
     "agents/codebase-mapper.md": "e1500018c3c67c8408f6b9fbc2543221d69c2884083d14327fe27e751fc920bc",
-    "agents/crm-pipeline-implementer.md": "bf097760f52a7d30e329c8e3049676efebfd677d40a29e05e0e961e2128e4d1f",
-    "agents/debugger.md": "bc9bf46863af9028920780f21894778f3c07f1f02dc2ad84246520f54d33a6f0",
-    "agents/evolution-go-integrator.md": "841b0d1fa1c962def8d47f8dc159a21a896a7fa6d7888eb42b6827d0bcab53e6",
+    "agents/crm-pipeline-implementer.md": "bf69eebbad5f35b7eb733fadf90b4655a6a3b72913cea3aa9f6b6902e9f1fa04",
+    "agents/debugger.md": "57d55aa17e96caba5294a632d56ed3d0c4952032ec30d23ee3adba74dd7f7a46",
+    "agents/detector-tenant-quente.md": "850630f254422e39166a6c9e8f08164aae975ef93fd99759bac7acedd26cc5cd",
+    "agents/evolution-go-integrator.md": "8d45d81ee71a4655267959279bced1e38e94d93a83943f4572afbd8ca4c457ad",
     "agents/example-reviewer.md": "cad1dcd4cafffc73e96de653809fdc2a9a8b1d8dc51865efd4d1daaf4284755f",
-    "agents/executor.md": "bc3af696fd820d32b0c5742c3c27a5f0ae527ae1627fb967b549ac193b34a853",
+    "agents/executor.md": "f62f125d1c9523d61d2a2aec1534cadc9a2bae69a441b2662a2e4e1e34bbbdbf",
     "agents/golden-signals-instrumenter.md": "de85b79ea09157ccc37c2de0facc3e6d4b9dba087e780330bc78be9ddd447297",
     "agents/incident-investigator.md": "521fcd4add75d2d4668162d19373823a9826bc45f23afdaf813b19f34113082d",
     "agents/integration-checker.md": "15c2badc2a8b650b7229e9ba32f481b08370afd3b80f41c7ae1c4cda7652811e",
-    "agents/invite-flow-implementer.md": "8ed29134ef89b1464df618c5ee9e7bbc9e2a28d235f72614f4a9b6535c2a38ed",
+    "agents/invite-flow-implementer.md": "9590a8ac8a430cb1a98cdbb59370b1890a53aee78f279aeb4116c9f4e09be938",
     "agents/legacy-characterizer.md": "1e9da6e7a0518a19788052e60554740f582fa2cd27b8dc1dc73251df3e712714",
-    "agents/lgpd-compliance-auditor.md": "ddde11f1867340b7e10c336696157538d1666df97997fcedc4dcaf997937fc56",
+    "agents/lgpd-compliance-auditor.md": "a34b1ddb76e6fcde0482fea309964f7a7450f2514745a23522e938d5181db83a",
     "agents/load-shedding-instrumenter.md": "63db224e3d8033a18cfe47ec8fc631685e356bf456d98b35c9c5d42d80fc5630",
-    "agents/multi-tenant-isolation-auditor.md": "eb4546c339469b5c9928786071a28807db38ec3e18213001e12a58e179938b2d",
-    "agents/multi-tenant-rls-writer.md": "cd23e46cf5d00ed09bf1a1a1f51f517d75531d1a9078d46239eab4abcd67de20",
+    "agents/multi-tenant-isolation-auditor.md": "b4c03df48bb563711922880144c2a3a71139559455eeb2fd444df77ee9e3cdc0",
+    "agents/multi-tenant-rls-writer.md": "74b1af763125e77e54976be9544db60bb8265eccf40429a8b29e9a24aa60fbbc",
     "agents/nyquist-auditor.md": "1d7590f356714eaacbdc92831dd100f9ca230a9c461e34223eb4bdf67ebaf076",
     "agents/observability-coverage-auditor.md": "e2c68e145182446ec47753b63063fea1d12b4356b390f33dffd4890f2c1a4352",
     "agents/observability-instrumenter.md": "6d40d96fdc3281b85d7db6d2e82f0e018945f2d148ab1b39c951774e6873aad2",
     "agents/omm-auditor.md": "8e09dca83495ba869f0f4c79c156f909a22b272e9ed77a0fe8a6658887cff7c0",
-    "agents/org-onboarding-implementer.md": "c63dc139be4b03db2019c83f75a21c37b2005eebd1638012b4a5b3f6ed9b60cf",
+    "agents/org-onboarding-implementer.md": "4887c62e5b0f235bff81ee008f481732d493356cad08f7905172acc1901e2cbb",
     "agents/payload-capture-instrumenter.md": "f1517b1a5d5cb10f8229de7402d4b6a786e8bd0ac9f6de3701a0a9dcebeb6cd5",
     "agents/phase-researcher.md": "3e431d8d6bd4f1459b049771a30f7176cfb6ccc21bf49e8a4cfb0bb03f9e7f7d",
     "agents/plan-checker.md": "32982f3713b4251d123e981c3f723c2a8702625edbbba4d862d64155086b94ae",
-    "agents/planner.md": "0cd395e7d235c8d103005a2e58a81d7f5dc2e9807d355234a7ddf8f3bf148c8e",
+    "agents/planner.md": "395ddace0a708de89a3e2fb66bbf36df706f172bdc2a66023012d056b9c30ac4",
     "agents/postmortem-writer.md": "d8c30abd537920bfc7b060a7b1e4ba98e8c0ca59ad4297e89eee3d35992f9509",
     "agents/project-researcher.md": "3cf77520ff888cf0a4c879e147180f6f9c3ba04bd9b0def10ef8f35e7ccedbf6",
     "agents/prr-conductor.md": "22710f35f702f132a9cdd5c6e3fe2dcf44cda24af36a2cd91060799dd19e9a02",
@@ -48,19 +50,24 @@
     "agents/shotgun-surgery-detector.md": "334966916715c4c5eff3324b81a09e17b312defd9a8549f17bfe19c06dbd9e92",
     "agents/slo-engineer.md": "1e6ec6e3031a40be6a6d631400f0dd12bb2b94ce37a9816d89d1badae1e7d7ca",
     "agents/storytelling-analyst.md": "dd68f9e8cb01bbd102acb9a2021be81a0b8d6f2528f30971f9aa5ef331f7e563",
-    "agents/supabase-architect.md": "6c4ff6be402da823ea8f4d5eabbaa30cbbb4d720da9257f10f88bbaa0b750cda",
-    "agents/supabase-auth-bootstrapper.md": "91a1f2d7df10cebf855e53364afc248a23b222dba2522d3084c9db0dfa5747f3",
+    "agents/supabase-architect.md": "523247c1ae7f69158d42971b19d51d299bc94a386ed871bdfe7eefbf3471a076",
+    "agents/supabase-auth-bootstrapper.md": "a197c3cd2c817f67a53400d5ba4fbf491f365a45e601fc60cd708ed07340272b",
+    "agents/supabase-column-privileges-writer.md": "dc6ed23511008cda6b9698afc3b577512547aaaf555d2a4de9273417f247ee6b",
     "agents/supabase-edge-fn-writer.md": "7b470664173f8c5c9ab12af74f58ddaeee3eb46bb3d2ca8de21b1c284961d9dd",
-    "agents/supabase-migration-writer.md": "4c0acc67bbf789a674453b12ed2aec73cb479609060c0db5349bf302544c2a75",
+    "agents/supabase-migration-writer.md": "a964dd676362184fe657446415202e0c620b0d4ce26ebe25125b6a55937c2fba",
+    "agents/supabase-rbac-implementer.md": "605de2d2a43a3d1eda2e181b98f53aae45b0c80123acc059305b65adf51f9f74",
     "agents/supabase-realtime-implementer.md": "e10d7f723734da0dd930d6e0481e2afb2abde5b470d9c45e4364889dac19f3d6",
-    "agents/supabase-rls-writer.md": "0ac667ba0f6543699b0053e5e8ec3eca43aff6e3307adde3959e9ce2056a9136",
+    "agents/supabase-rls-hardener.md": "2d20a112d9fac3e2d6f7012acfd1aa141608952bfe0412878dad0201411c1a56",
+    "agents/supabase-rls-writer.md": "d631cc957020f20612589b2b832c52d598c1a6eb70a263117071d9e0212165fb",
+    "agents/supabase-roles-implementer.md": "57936bdb20aca00177efe2db9043205ba266809d5d1dd5ea50675784dbe8f08c",
     "agents/supabase-storage-implementer.md": "28d57bc1750acb5b0624ecc33bac6e7855e16dec40df4c490865df885a0980f8",
-    "agents/super-admin-implementer.md": "dd3d7897396d96ad3a1a91e43128f1fac26c010b8974b08f1d0ddc6cf2e831c3",
+    "agents/super-admin-implementer.md": "2cc926adb3f0deb23c319fd301dd187e68be5edd095ca5d8d3b5fa9e1054ae57",
     "agents/toil-auditor.md": "58770f11805e7b3d8cc7b70d79ed73a6a33dc7d17b66fdc7716454406acefd61",
     "agents/ui-auditor.md": "a94816e9535757b02c6fcc8ae1e51f6378813c9f256e26859e01757e49c38d31",
     "agents/ui-checker.md": "be3308db8733d8f9ce3db1d2cef924738498ed03c7aa15c0fca21a9c15e79da2",
     "agents/ui-researcher.md": "9bdcb0d9b2ee007d98f13afd6932359e6450375e9c4a08f0d1ae72ad2ae349ac",
     "agents/user-profiler.md": "c6cc6a32afbc63992459447ee801ca7c77bd245a92615be110fc028274f41be8",
+    "agents/validador-evolucao-schema.md": "37f51c026a7a2860d225b4fdf84b61db0369c79e8ac16fab8745f69373b3ad97",
     "agents/verifier.md": "6d6f34947634981c11e7f1a0401a644668108e8d20e1e7c0992fb8c5ace58b77",
     "commands/adicionar-backlog.md": "ed36584b63e3d7bada207f30453b7995b2dbf0fe011889cc08b846bd069b1ca9",
     "commands/adicionar-fase.md": "5c321dd74bb36a4baef52c81ef2dcf1ef37898031bc55149a1a0aea9cc1b26f4",
@@ -84,6 +91,7 @@
     "commands/caracterizar.md": "994ce4136ba44b74890874f3274c26bcdc9f4feb5f4852cb0288687142ab1403",
     "commands/concluir-marco.md": "bd8c8ad217af6e4ab0487d35d98fae94a9ec5a5ffd0a2199871e0d484cefd438",
     "commands/configuracoes.md": "7657ed72bca4e7eba30ac74a08d8eeca2136aefdcedf18e739e6a89b2417a8c0",
+    "commands/dados-distribuidos.md": "b76977e87a795df24d82d59495401b52812da1e738d6b2545aabde756af83132",
     "commands/definir-perfil.md": "d1b73e411b94dad257bb3bac7ffc7f70deb92910442f8762f863076201b81f9b",
     "commands/definir-slo.md": "2f891393b6b46bffd28264d6cc340baf74297c7cd58be137d67c92313e942480",
     "commands/depurar.md": "2d69a0c3b0b375a20a4ef9f63a67a844d2f6c33f86f95cb603f4cb0cb5b2d362",
@@ -145,7 +153,7 @@
     "commands/setup-notion.md": "d1cd970ddecce7211c2cec1a42f0b5c74ea537b4f5e498924d52813542508ae9",
     "commands/sre.md": "eb4701d8d5fd98671ef562eddc4f021fce6ca5474cd840bb06df5b81f4741c2d",
     "commands/storytelling.md": "84b3344392e4562d9b92da86be70cf4d4804ca75fbdfcb11265511be801f390e",
-    "commands/supabase.md": "bda09c62f1c1d1ce5c339f0f4421347abffb86591c1af49880a1f2ac437a2c6b",
+    "commands/supabase.md": "136286e6d127d50963c989d60031907471f5826d5f7386cbad3c3a40aa6e5ecc",
     "commands/sync-main.md": "6a62ec23e0d4a9bf4b68f531fb65caad67583ce4bb73bcf8f63945dc793b664c",
     "commands/validar-fase.md": "43f18fb86d86e3b685581c242fc63169186725e2077cdacad047ca1fdfbcfb61",
     "commands/verificar-tarefas.md": "4e8af66691b1fd4895db1dea8a0eabd8f8e46682e259dbf9a8c5f6e1687c1a06",
@@ -292,22 +300,27 @@
     "hooks/statusline.js": "65c0113331f6de219e2224cb2403fa7e4e9202441895e48797732b20b0f2b48b",
     "hooks/workflow-guard.js": "197f1921c971236174f2f7cd65ae51e36ce5ac0ecbeac25c518ca379525fb0f6",
     "settings.json": "39e3e9bd40bc30a00c1790071462651e8ba6e52f5f16455a8f7bbacd7bde1827",
+    "skills/_shared-dados-distribuidos/glossary.md": "13d5cdb670411cad9b8946bb5fa3cac94dd613301295f37ca308984fcf2e42d0",
     "skills/_shared-legacy/glossary.md": "7dabc31ddd3b2b597f76d1a55ef51b0bd59315aaef59245873c224c4fd21c86d",
     "skills/_shared-multi-tenant/glossary.md": "1e040a36025489859312430771dde13bde9c62098fcd100440a82a2bb4d22b6a",
     "skills/_shared-observability/glossary.md": "ec3892c226af03299c0875e36fd0170cc9f801b02df52a2e0ec5c7468229912a",
     "skills/_shared-sre/glossary.md": "55a052c7d2292622150ed1cbb5aa0d675c332287b00ee4e3dd84900f9cf0ec84",
-    "skills/_shared-supabase/glossary.md": "2ebb4e09d9eda88a4f388f406f5cdb36fafa26a3ce6fb33d5c1976bcfac19327",
+    "skills/_shared-supabase/glossary.md": "24a794d195bdcf7ee6a5cc5c81b0ca93e3c8470afee6965fb104d9490072c450",
     "skills/ai-prompt-characterization/SKILL.md": "1a8114296c754e2018b1c1fd428c364f8de4485fedd5df78d3afcb33c3fef1a4",
-    "skills/audit-log-multi-tenant/SKILL.md": "8a7bc9b35f203098049a34b4dd5ec9693800ee1d8bf52f9c2bf0a78583c84323",
+    "skills/armadilhas-sistemas-distribuidos/SKILL.md": "fd33913c41a03f37eaa5fbc2aa3c33321397daa98a505ccb5f2d692dd8a00d5a",
+    "skills/audit-log-multi-tenant/SKILL.md": "877871e609008ec04c805c3f1c6de494c80adc838285eae119a23355a13bbe48",
     "skills/b2b-saas-architecture/SKILL.md": "01f39621cb924a885b3a097e9012d6dba637fcd9e543cd82505261190c227989",
     "skills/blameless-postmortems/SKILL.md": "55c386c055c869474b6164c0b871a882369a5e57ddf7de8eddcd881ceba70dfa",
     "skills/burn-rate-alerting/SKILL.md": "19cbd991efbb3b5d9ab6c53ada448b2624e6ab3f87ac500c0787bc4c8f2988e3",
-    "skills/cascading-failures/SKILL.md": "d308982fea601e8b9fa3ae9ec4bb7052475b57b039323f7398a956199f5edcd4",
+    "skills/cascading-failures/SKILL.md": "ec1f943310d8bbc3d22d2ba0499cc3e2c3123e2a0a90e07aa6814fbdaff4b0ee",
+    "skills/consistencia-leitura-replica/SKILL.md": "b65152e51d0c5a61bc0da3072d3186f17ea3bbeae53ddaa8de74e83dec5cfa11",
     "skills/core-analysis-loop/SKILL.md": "298f27d1706abe0e64fa7dfa925eba3255328dccd03b24589be1e80d6d673dd3",
-    "skills/crm-lead-pipeline-patterns/SKILL.md": "15ec4522d1c4f0a571fdc9db989552c5f72b1880ecbc4ab538435284f06d75b7",
+    "skills/crm-lead-pipeline-patterns/SKILL.md": "6e04e34cf901e33412e6909be59555bb014079b201223f06a1fa32b34c649c4f",
     "skills/distributed-tracing/SKILL.md": "7747bbd0563b28cba11b11064012c0b7a696ac04bd65c64debb42987781b26ff",
     "skills/eliminating-toil/SKILL.md": "451c7dfc5dd5af925f0b38657553ddbc1cc95923b3a572ac79d18201acb136cd",
+    "skills/escolha-modelo-consistencia/SKILL.md": "c32a39848fbaaa3ea4e4e15cbc09146fa5283bdbfc616428516ccbd9bd6ae741",
     "skills/event-based-slos/SKILL.md": "6bf107d86cc409ab24773e16f26b4eca6e228dfd2db027de7b384e88a05bd50e",
+    "skills/evolucao-schema-compativel/SKILL.md": "2082520c4b86787b039d8373b7076ed857bd843bb7f83d6721bb375091167f60",
     "skills/evolution-go-whatsapp-integration/SKILL.md": "4b964cb297ec9df7413b1cc6dd73fa3d04a8de68c584b9b72fceed576d2afebe",
     "skills/example-skill/SKILL.md": "2ddf878e69e19b0138d8a2c6a31404593fed3e229375869991ab28b469069233",
     "skills/four-golden-signals/SKILL.md": "c83bac06bb676a3a7ee24c11d9d5d72388ae8bfd21ad9e7face24c5fcebd7b60",
@@ -327,35 +340,42 @@
     "skills/load-shedding-graceful-degradation/SKILL.md": "b261e4ff756e4aea689ccfd3c4dc721df22c6ae91b08a029db2b7a9f2a172b6f",
     "skills/member-invite-flow/SKILL.md": "caea28f2a6c63f585f92b1fdbd113818a6f41aedef5ee9482bd60ea881f660fc",
     "skills/member-management-react-shadcn/SKILL.md": "232fa16b9f3def2d4df46cefdcf1f82ce7db211baf3d5f8e5255ce58f21351bb",
-    "skills/multi-tenant-performance-scaling/SKILL.md": "ab38dcae7d99f7cbd2c0e38fcb388d413493e28de18983d648f6d9b6ed47fc3c",
-    "skills/multi-tenant-rls-hierarchy/SKILL.md": "a48f60272d5c5bc9be84596a9b245651f1aec91cfa87e40ea1404e57ac8c54b5",
+    "skills/multi-tenant-performance-scaling/SKILL.md": "ca31b2f421c27158119e467cddf71f1de8905cec632de8948703f13344621895",
+    "skills/multi-tenant-rls-hierarchy/SKILL.md": "c1e31d6a0bc27cba3520125dc0bc0266a81e9e1bc86d35a04f553fd5141a8b94",
     "skills/observability-driven-development/SKILL.md": "9abaec925bbb59391597d92164c43b432eef32884ce3b4623f419f421322fa83",
     "skills/observability-maturity-model/SKILL.md": "49bccaa9e122651fbbe5457f7fef96114f9b3e8626faf6e3a8cc6ded66c7da10",
     "skills/opentelemetry-standard/SKILL.md": "891099b62991c74598f1e1c51889475cd69ca624f92bd0e7931b2613e4f90f01",
     "skills/org-onboarding-flow/SKILL.md": "68be27185492a93a599860a7212e935d4a6e3625ffd86348d18086657c04fa0b",
     "skills/org-switcher-react-pattern/SKILL.md": "c246d536d1ee65a9ea0cc7ab09429ca798afb7ef3468eaaeacedc0bd450bb963",
     "skills/permission-gate-react-pattern/SKILL.md": "2c3f8b8ac0c495af0991c5e314d962187ef0cf61982440f0959b422385b07969",
+    "skills/postgres-isolamento-concorrencia/SKILL.md": "7398de91667f7701dd47d0f03800a2fd97ac15e9b13c52be9ac9171f9cf3d8f5",
     "skills/pre-refactor-characterization/SKILL.md": "9124f9ca0636a75474ea3f6d851e587be2f75505b3a835af0a4aaa0855bd20d3",
     "skills/production-readiness-review/SKILL.md": "2a9731265163c9fe7ba4fd05ceaf164ee4d1188b0d147ddff3b13bd9d3058c04",
-    "skills/rbac-permissions-matrix-supabase/SKILL.md": "7cf6aafe6d0de895165fb67649d54147341a4f103f13624c8981fe16d2c7d2c9",
+    "skills/rbac-permissions-matrix-supabase/SKILL.md": "8c0bf2b2f7935fdb3024954e797ac11604457a4646a61dc01eb06a985ee1eabd",
     "skills/release-engineering/SKILL.md": "01e69f50d2bb207d348552a01d0d69b6159b47573fe7e31aec53f6df52c3d057",
     "skills/retry-strategies/SKILL.md": "017a38146787592cde5c009bc06c8f483ca2b609a018d0b526972ddf5e46f52a",
     "skills/sre-risk-management/SKILL.md": "6e56a30b081abffbf9ce97e86b9c376361d6af765fe5475970f1646351c54e39",
+    "skills/streams-eventos-cdc/SKILL.md": "64af5564e999ec0d8c3a0e1edbbea9f80a212d8ce8d03b0dfcf9c2db29b76d8f",
     "skills/structured-events/SKILL.md": "a693c8a19709066ea60860a01ba54731406d7daf41ed51adea9c29a2de131fac",
     "skills/supabase-auth-ssr/SKILL.md": "941d80ad88b4cbeccadf852d82f64f0167bce204005f72b32bc2aaf81a460af6",
-    "skills/supabase-cron-queues/SKILL.md": "e7abbb57de2f49a3feefe8d160b4bc1c3e00629fa824755872d7b6cb84eba5d0",
-    "skills/supabase-database-functions/SKILL.md": "9eaf17a5b75f3e8c398211f032a939fa4f7517c0453e977d84bf364f39cf550d",
+    "skills/supabase-column-level-security/SKILL.md": "0fcdac70be44ffbdccb56d5b0ca4f80c1267ddba037a757fcf969013754e521f",
+    "skills/supabase-cron-queues/SKILL.md": "4f48ed9cea9b5b2bc187983ffbbb63f9e46ad65ff6a6306b5c8b4b01b4e6911a",
+    "skills/supabase-custom-claims-rbac/SKILL.md": "2120bd49e0b6f1ee9723128654550836634f93dee5a80e6c6409fe96af223a0c",
+    "skills/supabase-database-functions/SKILL.md": "77b49f2930d61667e4ae1839944dbee012267253444aef63cc1dc24d227deff7",
     "skills/supabase-declarative-schema/SKILL.md": "8a78cae2d74287002c02bafdfb8218a9ac20b7d75047c269c702d9b8e3d22476",
     "skills/supabase-edge-functions/SKILL.md": "bf195e3fbce2bd94cb782ce15ecc60260217ac40d9ac5cbc787362de6629f960",
-    "skills/supabase-migrations/SKILL.md": "bd502f715d816187b45ae34e400c47ea548fe0f6c077813ce99f2d451ac3db7c",
+    "skills/supabase-migrations/SKILL.md": "bd7a4e43e2c135d0f15f3ac4045c22e378b863b2c878c1895486cdfb39f4e968",
     "skills/supabase-pgvector-rag/SKILL.md": "cd50663c5b19d08d9bc17bc9b4444f7fc2f6910f5c52502e7c50b1578ebe7e70",
+    "skills/supabase-postgres-roles/SKILL.md": "eec741701e1f71b380c0275c318968d0674e20017437a00e39808b33115e62fa",
     "skills/supabase-postgres-style/SKILL.md": "4e48bd0a9ed46bea7c3be97ef749e5c148369ceca08ef3dc8d813d8a03a48703",
     "skills/supabase-realtime/SKILL.md": "ca2584a59742b30f5351fad23f4a1957218ca730ce3af990affe79f03854f460",
-    "skills/supabase-rls-policies/SKILL.md": "b8cab2e5813a00fea6aa19a59be94dfa536d675067c2e87c94576e97d472d16e",
+    "skills/supabase-rls-defense-in-depth/SKILL.md": "8881cac68fd72cf162dfc765baaffb0fa5d0f282d730792cdc05ab602cbe5efd",
+    "skills/supabase-rls-policies/SKILL.md": "eae0bc3ae5e775e72c767de9f721b87a044613e801304dad2389c33dd620995d",
     "skills/supabase-storage/SKILL.md": "f7360aa9149e55f68fa794a91c18994329e4f304cc263f90f0607e43053e9da8",
-    "skills/super-admin-platform-pattern/SKILL.md": "e3063f32f111756fbc1792767b2a26a9aef2ed16b90615cc21655acc33269ae7",
+    "skills/super-admin-platform-pattern/SKILL.md": "aec4c25fd8f8314e6cd5b45037a56ccad5ef1599f09daa70d089b92ef2ac28df",
     "skills/telemetry-pipelines/SKILL.md": "7623244afdf8e6b0b865e572c8e8537c73255914a4562a95f99f22be7448f80e",
     "skills/telemetry-sampling/SKILL.md": "d448f9022ba0f53a9f692189d2515e45fc3e9de52256bd0715b4c1374c8cc203",
+    "skills/tenant-quente-mitigacao/SKILL.md": "97bcca2e8fad4fb3aa3d01d604720ed7cb5c5fcaee6bff8780d6ce4904dac5c1",
     "skills/whatsapp-conversation-state-machine/SKILL.md": "8467327173d17213406e2ae2c959984e60399eee567384bc39a4592f7f09f747"
   }
 }