Explica código, projetos, arquitetura ou qualquer parte do sistema em linguagem simples, sem jargão corporativo, de forma objetiva e com diagramas (Mermaid). Dispare quando o usuário pedir para explicar, entender ou detalhar código/projeto/arquitetura de forma simples ou didática. Trigger em: "explica esse código", "explica esse projeto", "como funciona", "como isso funciona", "explica a arquitetura", "quero entender", "explica de forma simples", "explica sem jargão", "faz um diagrama disso", "faz um fluxograma disso", "explain this code", "explain the architecture", "how does this work", "walk me through this".
Testa rotas de API e faz varredura de segurança (DAST leve) contra alvos autorizados. Use para testar rotas, route testing, exercitar endpoints via curl, mapear rotas de um backend (Express/Fastify/Nest, Flask/FastAPI/Django, Go, Spring, Rails, Laravel, .NET, actix-web e outros), rodar matriz de auth testing e access control, detectar vulnerabilidades — SQLi, NoSQLi, XSS, command injection, path traversal — via fuzzing e injection em nível de detecção, gerar coleções Insomnia/Postman v2.1/OpenAPI (swagger), e produzir relatório com achados de segurança e sugestões. Dispare quando o usuário mencionar: testar rotas, route testing, curl endpoints, DAST, fuzzing, injection, SQLi, XSS, auth testing, access control, insomnia, postman, openapi, swagger, vulnerabilidade, coleção de rotas, testar minha API, varredura de segurança, pentest da própria aplicação. Agnóstica de linguagem e framework: tudo específico do alvo é declarado em runtime.
Creates a Kanban-style TASK-PLAN.md with CARD-XXX cards from a free-text task description. Each card has Description, Scope, Acceptance Criteria, Dependencies, and Checklist. Use when the user describes a feature or task and wants a structured Kanban plan, card decomposition, TASK-PLAN.md, structured cards, kanban planning, or runs /kanban-task-planner.
Executes a TASK-PLAN.md created by /kanban-task-planner. Reads the plan, executes each card's checklist directly in the conversation (no subagents), respects dependency groups, and marks cards complete as they finish. Use when the user wants to run a kanban plan, execute TASK-PLAN.md, run cards, or runs /run-kanban.
Creates a new skill+agent combination from a workflow description. Use when the user wants to automate a new workflow, create a new slash command, build a new subagent, or design a skill+agent architecture for a recurring task. Trigger on phrases like "cria uma skill", "cria um agent", "cria skill+agent", "quero automatizar", "novo workflow", "nova skill", "novo agent", "create skill", "create agent", "build agent", "build skill", "new workflow".
Transforms a raw development request into a structured task.md spec. Reads developer/requests/[name].md or an inline description, consults the architecture-guardian, and creates developer/tasks/[name]/task.md with full scope definition.
Orchestrates parallel execution of all tasks in a plan. Reads developer/tasks/[name]/MAIN.md, launches task-executor agents respecting dependencies, verifies each task with task-verifier, and runs the project's type-check at the end.
Generates NatSpec-style documentation comments for complex implementations in any language or framework. Use when the user wants to add, complete, or audit documentation comments. Works with TypeScript, JavaScript, Solidity, Python, Rust, Go, and any other language. Triggers on phrases like "natspec", "adicionar natspec", "documentar função", "documentar classe", "gerar docs", "comentários natspec", "missing docs", "document this", "add comments", "adicionar comentários", "documentar implementação", "generate natspec", "@notice", "@param", "@return", "@dev".