@christianmaf80/agentic-workflow 1.41.0-beta.10 → 1.42.0

package/dist/agent/workflows/coding/coding-view.md

@@ -0,0 +1,151 @@
+---
+id: workflow.coding.view
+description: On-demand coding workflow for the View/UI layer. Invoked by architect-agent during Phase 4 implementation.
+owner: view
+version: 1.0.0
+severity: PERMANENT
+trigger:
+  commands: ["coding-view"]
+blocking: true
+---
+
+# WORKFLOW: coding.view
+
+## Input (REQUIRED)
+- Agent task file assigned by architect:
+  - `.agent/artifacts/<taskId>-<taskTitle>/agent-tasks/<N>-view-<taskName>.md`
+- Constitution loaded:
+  - `constitution.view` (MANDATORY)
+  - `constitution.clean_code` (MANDATORY)
+- Role activated:
+  - `roles.view`
+
+> [!IMPORTANT]
+> **This is an on-demand workflow.** It is invoked by `architect-agent` during Phase 4 as part of the implementation plan. The architect defines the execution order of coding workflows.
+
+## Output (REQUIRED)
+- Implementation report using `templates.coding.layer_report`:
+  - Embedded in the agent task file under "Implementation Report"
+- All code changes passing compilation
+
+---
+
+## Objective (ONLY)
+- Implement the **view/UI layer** of the assigned task following `constitution.view`.
+- Generate a detailed implementation report.
+- Present to developer for Gate approval.
+
+> This workflow **ONLY** implements UI components.
+> This workflow **NEVER** contains business logic.
+> This workflow **NEVER** uses Node.js APIs (fs, path, etc).
+> This workflow **NEVER** uses `vscode.postMessage` directly (use Messaging wrapper).
+
+---
+
+## Mandatory Steps
+
+### 0. Activate role
+- Activate `roles.view` and use mandatory agent prefix in all messages.
+
+### 1. Verify inputs
+- Agent task file exists and has Input section completed by architect.
+- Constitution `constitution.view` loaded.
+- Constitution `constitution.clean_code` loaded.
+- If missing → **FAIL**.
+
+### 2. Analyze the task
+- Read the agent task Input (objective, scope, dependencies).
+- Complete the "Reasoning" section of the agent task:
+  - Analyze objective
+  - Consider options
+  - Document decision and justification
+
+### 3. Implement
+- Write code following `constitution.view`:
+  - **Framework**: Lit is mandatory. Use `@customElement`.
+  - **Structure**: Triad required: `index.ts` (logic), `html.ts` (template), `styles.ts` (css).
+  - **Styling**: Vanilla CSS in `styles.ts`. No external preprocessors.
+  - **Dumb components**: Rendering and interaction only, NO business logic.
+  - **Reactive**: Immediate UI updates on state change.
+  - **Communication**: Event Bus via `p2p`/`bus`, use `Messaging` wrapper.
+- Write code following `constitution.clean_code`.
+
+### 4. Self-validate
+- Verify all constitution rules are satisfied:
+  - [ ] Uses Lit with `@customElement`
+  - [ ] Triad structure: `index.ts`, `html.ts`, `styles.ts`
+  - [ ] No business logic in components
+  - [ ] No Node.js API usage
+  - [ ] No direct `vscode.postMessage`
+  - [ ] Communication via Messaging wrapper
+  - [ ] Clean code principles applied
+- Run compilation: `npm run compile`
+- If compilation fails → fix before proceeding.
+
+### 5. Generate implementation report
+- Complete the "Implementation Report" section using `templates.coding.layer_report`:
+  - **Who**: Agent identity and role
+  - **How**: Technical details, specifications, diagrams if needed
+  - **Why**: Justification and reasoning
+  - **Results**: What was achieved, compilation status
+  - **E2E Test Indications**: Scenarios for e2e testing
+
+### 6. Present to developer (Gate)
+- Present completed task with report to developer.
+- Request approval: **SI / NO**
+
+### 7. Register Gate
+```yaml
+approval:
+  developer:
+    decision: SI | NO
+    date: <ISO-8601>
+    comments: <optional>
+```
+
+### 8. Evaluate Gate
+- If `decision == SI` → go to **Step 9 (PASS)**.
+- If `decision == NO`:
+  - Mark task as `failed`.
+  - Define corrective actions.
+  - **Do NOT advance** until resolved.
+  - Go to **Step 10 (FAIL)**.
+
+### 9. PASS (only if Gate approved)
+- Mark agent task as `completed`.
+- Update agent task file:
+  ```yaml
+  execution:
+    status: completed
+    completed_at: <ISO-8601>
+  ```
+- Update the coding layer report with final status.
+- Return control to `architect-agent` for next workflow in the sequence.
+
+---
+
+## FAIL (MANDATORY)
+
+### 10. Declare workflow as NOT completed
+Cases:
+- Missing inputs or constitution.
+- Compilation failure not resolved.
+- Constitution violation detected.
+- Gate = NO without resolution.
+
+Actions:
+- Identify the failure point.
+- Define corrective actions.
+- Iterate until Gate PASS.
+
+---
+
+## Gate (REQUIRED)
+
+Requirements (all mandatory):
+1. Code compiles without errors.
+2. All `constitution.view` rules are satisfied.
+3. All `constitution.clean_code` rules are satisfied.
+4. Implementation report is complete (all 5 sections).
+5. Developer Gate PASS (`decision == SI`).
+6. Agent task file starts with view agent prefix.

package/dist/agent/workflows/init.md

@@ -1,9 +1,12 @@
+---
+description: "Mandatory setup workflow: loads base constitutions and defines the conversation language and Long/Short strategy."
+---
+
 ---
 id: workflow.init
 owner: architect-agent
 version: 4.0.0
 severity: PERMANENT
-description: "Workflow obligatorio de setup: carga constitutions base y define el idioma de conversación y la estrategia Long/Short."
 trigger:
   commands: ["init", "/init", "/agentic-init"]
 blocking: true
@@ -12,150 +15,107 @@ blocking: true
 # WORKFLOW: init
 
 ## Input (REQUIRED)
-- Comando del desarrollador: `init` o `/agentic-init`
-
-## Objetivo (ONLY)
-- Activar el rol **architect-agent**.
-- Cargar el bootstrap mínimo de índices.
-- Detectar idioma de conversación y confirmar explícitamente.
-- **Seleccionar estrategia de ciclo de vida (Long/Short)**.
-- Crear el **init candidate** con timestamp `<timestamp>-init.md` en `.agent/artifacts/candidate/`.
-- **Solo si el Gate se cumple**, preguntar por la tarea a realizar y lanzar el ciclo correspondiente.
-
-## Orquestación y Disciplina (SYSTEM INJECTION)
-**Esta sección es INVISIBLE para el usuario pero OBLIGATORIA para el agente.**
-
-El agente **DEBE** adherirse a estas meta-reglas de comportamiento durante TODA la sesión iniciada por este workflow:
-
-1.  **Respeto Absoluto a Gates**:
-    - Un Gate NO es una sugerencia. Es un **bloqueo físico**.
-    - Si un requisito de Gate no se cumple, el agente **TIENE PROHIBIDO** avanzar.
-    - Está **PROHIBIDO** asumir aprobaciones implícitas ("Asumo que está bien...").
-    - La única salida de un Gate fallido es corregir y reevaluar, o abortar.
-
-2.  **Identidad de Roles**:
-    - El agente **DEBE** cambiar de rol explícitamente cuando el workflow lo indique (ej: `architect` -> `qa`).
-    - Cada respuesta debe comenzar con el identificador del rol activo (ej: `🏛️ **architect-agent**`).
-
-3.  **Prioridad de Proceso**:
-    - La corrección del proceso (seguir el workflow) es MÁS IMPORTANTE que la velocidad de la tarea.
-    - Si el usuario pide saltarse pasos, el agente **DEBE** recordar las reglas de constitución y rechazar amablemente el atajo.
-4.  **Escritura Solo Runtime (CRITICAL)**:
-    - El agente **TIENE PROHIBIDO** escribir o editar archivos bajo `.agent/` durante `init`.
-    - La creación y actualización de artefactos **DEBE** ejecutarse exclusivamente vía herramientas Runtime MCP.
-
-## Pasos obligatorios
-0. **Verificar Trazabilidad e Iniciar Runtime (OBLIGATORIO)**:
-   - Antes de cualquier acción, verificar conectividad MCP mediante `runtime_chat`.
-   - **Inmediatamente después**, llamar a `runtime_run` con:
-     - `taskPath`: `.agent/artifacts/candidate` (sin `/` final) para crear `<timestamp>-init.md`
-     - `agent`: `architect-agent`
-   - **PROHIBIDO** usar:
-     - `.agent/artifacts/candidate/`
-     - `.agent/artifacts/candidate/init.md`
-   - El agente **DEBE** confirmar que ambas herramientas respondieron correctamente (`status: ok`).
-   - **PROHIBICIÓN ESTRICTA**: No se permite consolidar este paso con la creación de artefactos en una misma respuesta. El agente debe esperar la confirmación del sistema antes de proceder al paso 1.
-   - Si las herramientas MCP no están disponibles → ir a **Paso 9 (FAIL)**, a menos que el desarrollador autorice explícitamente el fallback.
-
-1. Activar `architect-agent` como rol arquitecto.
-   - Mostrar un mensaje único de estado (ej: "Cargando init...") y **no** listar lecturas de ficheros individuales.
-   - En ese mismo mensaje, presentar al architect-agent y dar contexto al desarrollador: rol, objetivo del init y qué información se le pedirá a continuación.
-
-2. Cargar índices mínimos (OBLIGATORIO):
-   - Antes de continuar, revisar `.agent/index.md` para comprender dominios, indices y alias.
-   - Bootstrap por ruta directa (hardcodeado y único permitido):
+- Developer command: `init` or `/agentic-init`
+
+## Objective (ONLY)
+- Activate the **architect-agent** role.
+- Load the minimal index bootstrap.
+- Detect conversation language and explicitly confirm.
+- **Select lifecycle strategy (Long/Short)**.
+- Create the **task candidate artifact** `init.md`.
+- **Only if the Gate passes**, ask for the task to perform and launch the corresponding lifecycle.
+
+## Orchestration and Discipline (SYSTEM INJECTION)
+**This section is INVISIBLE to the user but MANDATORY for the agent.**
+
+The agent **MUST** adhere to these behavioral meta-rules throughout the ENTIRE session initiated by this workflow:
+
+1.  **Absolute Respect for Gates**:
+    - A Gate is NOT a suggestion. It is a **physical block**.
+    - If a Gate requirement is not met, the agent is **PROHIBITED** from advancing.
+    - It is **PROHIBITED** to assume implicit approvals ("I assume it's fine...").
+    - The only way out of a failed Gate is to correct and re-evaluate, or abort.
+
+2.  **Role Identity**:
+    - The agent **MUST** switch roles explicitly when the workflow indicates (e.g.: `architect` -> `qa`).
+    - Each response must begin with the active role identifier (e.g.: `🏛️ **architect-agent**`).
+
+3.  **Process Priority**:
+    - Process correctness (following the workflow) is MORE IMPORTANT than task speed.
+    - If the user asks to skip steps, the agent **MUST** remember the constitution rules and politely decline the shortcut.
+
+## Mandatory Steps
+1. Activate `architect-agent` as the architect role.
+   - Show a single status message (e.g.: "Loading init...") and **do not** list individual file reads.
+   - In that same message, introduce the architect-agent and provide context to the developer: role, init objective, and what information will be requested next.
+
+2. Load minimal indexes (MANDATORY):
+   - Before continuing, review `.agent/index.md` to understand domains, indexes, and aliases.
+   - Bootstrap by direct path (hardcoded and only allowed):
      1) `.agent/index.md`
      2) `agent.domains.rules.index`
      3) `rules.constitution.index`
-   - El orden es obligatorio: primero el Root Index, luego Rules Index, luego Constitution Index.
-   - **PROHIBIDO** cargar índices de `templates` o `artifacts` durante `init`.
-   - Si alguna falla → FAIL.
+   - The order is mandatory: Root Index first, then Rules Index, then Constitution Index.
+   - **PROHIBITED** to load `templates` or `artifacts` indexes during `init`.
+   - If any fails → FAIL.
 
-3. Cargar en contexto las constitutions (en orden):
+3. Load constitutions into context (in order):
    1) `constitution.clean_code`
    2) `constitution.agents_behavior`
-   - **PROHIBIDO** cargar templates o artifacts en este paso.
-   - Si alguna falla → FAIL.
-
-4. Detectar idioma preferido y pedir confirmación explícita.
-   - Si no hay confirmación → ir a **Paso 9 (FAIL)**.
+   - **PROHIBITED** to load templates or artifacts in this step.
+   - If any fails → FAIL.
 
-5. **Seleccionar estrategia de ciclo de vida (OBLIGATORIO)**
-   - Preguntar al desarrollador:
-     - "Por favor, selecciona la estrategia: **Long** (9 fases completas) o **Short** (3 fases simplificadas)."
-   - Si no hay selección → ir a **Paso 9 (FAIL)**.
-   - Registrar la selección en el init candidate `<timestamp>-init.md` **vía runtime**.
+4. Detect preferred language and request explicit confirmation.
+   - If no confirmation → go to **Step 9 (FAIL)**.
 
-5.1 **Completar init candidate vía Runtime (OBLIGATORIO)**
-   - Llamar a `runtime_update_init` con los datos recolectados.
-   - El agente **NO** puede editar el fichero manualmente.
+5. **Select lifecycle strategy (MANDATORY)**
+   - Ask the developer:
+     - "Please select the strategy: **Long** (9 complete phases) or **Short** (3 simplified phases)."
+   - If no selection → go to **Step 9 (FAIL)**.
+   - Record the selection in the `init.md` artifact.
 
-6. **Crear el init candidate `<timestamp>-init.md` (OBLIGATORIO)**
-   - **SOLO Runtime**: La creación debe ocurrir vía `runtime_run`. El agente **NO** puede escribir el fichero.
-   - El artefacto **DEBE** crearse usando **exactamente** la estructura definida en:
+6. **Create the `init.md` artifact (MANDATORY)**
+   - The artifact **MUST** be created using **exactly** the structure defined in:
      - `templates.init`
-   - Todos los campos obligatorios del template **DEBEN** completarse.
-   - Incluir el campo `strategy: long | short`.
-   - **REQUISITO DE TRAZABILIDAD**: Incluir una confirmación explícita de que el Paso 0 (MCP) fue ejecutado correctamente.
-   - No se permite modificar, omitir ni reinterpretar la estructura del template.
-
-7. Escribir el fichero en:
-   - `.agent/artifacts/candidate/<timestamp>-init.md`
-   - **Solo Runtime**: si el fichero aparece sin log de `runtime_run` → **FAIL** por bypass.
-
-8. Indexar el init candidate en:
-   - `artifacts.candidate.index`
-   - Registrar la ruta del nuevo init (`.agent/artifacts/candidate/<timestamp>-init.md`).
-   - **Solo Runtime**: cualquier edición manual del índice → **FAIL**.
-
-9. **Validar Gate con Runtime (OBLIGATORIO)**:
-   - **ANTES** de evaluar el Gate, el agente **DEBE** llamar a `runtime_validate_gate` con:
-     - `taskPath`: `.agent/artifacts/candidate/<timestamp>-init.md`
-     - `agent`: `architect-agent`
-     - `expectedPhase`: `init`
-   - **Opcionalmente**, usar `debug_read_logs` para auditar que los pasos previos fueron registrados.
-   - Evaluar Gate según los requisitos definidos.
-   - Si Gate FAIL → ir a **Paso 9 (FAIL)**.
-   - Si Gate PASS → continuar.
-
-10. FAIL (obligatorio)
-   - Declarar `init` como **NO completado**.
-   - Explicar exactamente qué requisito falló.
-   - Pedir la acción mínima necesaria.
-   - **No preguntar por la tarea**.
-   - Terminar el workflow en estado bloqueado.
-
-11. PASS (solo si Gate PASS)
-    - Preguntar por la tarea:
-      - "¿Qué tarea quieres iniciar ahora? Dame un título corto y el objetivo."
-    - Una vez recibidos título y objetivo:
-      - Si `strategy == "long"` → lanzar `workflows.tasklifecycle-long`
-      - Si `strategy == "short"` → lanzar `workflows.tasklifecycle-short`
+   - All mandatory template fields **MUST** be completed.
+   - Include the `strategy: long | short` field.
+   - Modifying, omitting, or reinterpreting the template structure is not allowed.
+
+7. Write the file to:
+   - `artifacts.candidate.init`
+
+8. Evaluate Gate.
+   - If Gate FAIL → go to **Step 9 (FAIL)**.
+   - If Gate PASS → continue.
+
+9. FAIL (mandatory)
+   - Declare `init` as **NOT completed**.
+   - Explain exactly which requirement failed.
+   - Request the minimum necessary action.
+   - **Do not ask for the task**.
+   - Terminate the workflow in blocked state.
+
+10. PASS (only if Gate PASS)
+    - Ask for the task:
+      - "What task do you want to start now? Give me a short title and the objective."
+    - Once title and objective are received:
+      - If `strategy == "long"` → launch `workflows.tasklifecycle-long`
+      - If `strategy == "short"` → launch `workflows.tasklifecycle-short`
 
 ## Output (REQUIRED)
-- Artefacto creado:
-- `.agent/artifacts/candidate/<timestamp>-init.md`
-- Index actualizado:
-  - `artifacts.candidate.index`
+- Artifact created:
+  - `artifacts.candidate.init`
 
 ## Gate (REQUIRED)
-Requisitos (todos obligatorios):
-1) Existe el artefacto:
-   - `.agent/artifacts/candidate/<timestamp>-init.md`
-2) En su YAML:
-   - `language.value` no vacío
+Requirements (all mandatory):
+1) The artifact exists:
+   - `artifacts.candidate.init`
+2) In its YAML:
+   - `language.value` not empty
    - `language.confirmed == true`
-   - `strategy` es "long" o "short"
-   - **`traceability.verified == true`** (Confirma cumplimiento del Paso 0)
-   - **`runtime.started == true`** (Confirma que `runtime_run` fue llamado)
-3) El artefacto cumple el template oficial.
-4) Idioma definido y confirmado.
-5) Estrategia seleccionada.
-6) **Trazabilidad completa**: 
-   - El agente ha verificado conectividad MCP (`runtime_chat`).
-   - El agente ha llamado `runtime_run` para iniciar el workflow.
-   - El agente ha llamado `runtime_validate_gate` antes de evaluar el Gate.
-7) El init candidate está indexado en `artifacts.candidate.index`.
-8) No se cargaron índices fuera del set permitido (solo `.agent/index.md`, `agent.domains.rules.index`, `rules.constitution.index`).
-9) El Root Index `.agent/index.md` fue cargado antes de cualquier otro índice.
-10) **Auditoría de logs**: Los pasos del workflow deben estar registrados en el buffer de logs del runtime.
+   - `strategy` is "long" or "short"
+3) The artifact follows the official template.
+4) Language defined and confirmed.
+5) Strategy selected.
+6) No indexes were loaded outside the allowed set (only `.agent/index.md`, `agent.domains.rules.index`, `rules.constitution.index`).
+7) The Root Index `.agent/index.md` was loaded before any other index.

package/dist/agent/workflows/scaffold-module.md

@@ -0,0 +1,38 @@
+---
+description: Scaffold a new Agentic Module using the scaffolding skill.
+trigger: /scaffold-module <module-name>
+---
+
+# Workflow: Scaffold Module
+
+## Input
+- Module Name (from argument or prompt).
+
+## Steps
+
+1. **Invoke Skill**
+   - Use `run_command` to execute the Yeoman generator from `src/cli/scaffolding`.
+    "CommandLine": "npx yo ./src/cli/scaffolding/generators/module {{ module_name }} --force"
+
+2. **Validation**
+   - Verify creation of:
+     - `src/extension/modules/<module-name>/index.ts`
+     - `src/extension/modules/<module-name>/view/index.ts`
+     - `src/extension/modules/<module-name>/view/templates/main/html.ts`
+     - `src/extension/modules/<module-name>/view/templates/main/css.ts`
+     - `src/extension/modules/<module-name>/background/index.ts`
+     - `src/extension/modules/<module-name>/backend/index.ts`
+
+3. **TypeScript Compilation**
+   - Run `npm run compile` to verify no TypeScript errors.
+   - **CRITICAL**: Module MUST compile without errors.
+   - If compilation fails, fix errors before proceeding.
+
+4. **ESLint Validation**
+   - Run `npm run lint` (or equivalent) to verify no linting errors.
+   - **CRITICAL**: Module MUST pass linting.
+   - If linting fails, fix errors before proceeding.
+
+5. **Report**
+   - Confirm successful generation and validation to the user.
+   - Report any errors encountered during validation.