playbook-forge

Exporta el ultimo playbook a una ruta especifica

Genera un playbook HTML completo con 13 flujos AI-native, 4 katas y adopcion progresiva

Ingesta archivos de contexto para enriquecer el proximo playbook

Abre el ultimo playbook generado en el browser

Genera una sola seccion del playbook (ej: timeline, katas, flujos, semaforo)

Shared configuration for all playbook-forge agents.

Use this agent when the playbook brief is ready and a complete content manifest needs to be generated. This agent architects all 13 flows, 5 katas (Shu-Ha-Ri), 65 interactive modals, bilingual content (ES/EN), 13 anti-patterns, and every section of the v2 playbook content. <example> Context: Generating content for a QA team playbook user: "Generate the content manifest for a GenAI playbook targeting QA leads who use Jira and Confluence." assistant: "Building the manifest with 13 flows adapted to QA workflows. Mapping Jira for task tracking, Confluence for documentation, and Drive->NLM->Gemini for the grounding layer..." <commentary>The strategist adapts generic flow templates to the user's specific tool stack and role context.</commentary> </example> <example> Context: Generating content for an operations team user: "Create manifest for IT Operations team that uses ServiceNow and Azure DevOps. Their main problems are incident response time and knowledge silos." assistant: "Adapting flows 1-4 to ServiceNow incident workflows. Flows 5-8 will cover Deep Research for root cause analysis. Architecture layers map to SharePoint->NLM->Gemini..." <commentary>The strategist maps the 3-layer architecture to the user's existing tooling and tailors exercises to real operational scenarios.</commentary> </example>

Use this agent when source files (meeting notes, process documents, discovery artifacts) need to be analyzed and converted into a structured context brief before playbook generation begins. <example> Context: User provides a directory full of meeting notes user: "Analyze the files in ./discovery/ and extract context for the playbook." assistant: "Scanning ./discovery/ ... Found 8 markdown files and 3 PDFs. Extracting roles, tools, pain points, and architecture patterns..." <commentary>The ingester scans broadly, reads each file, and produces a structured JSON brief that downstream agents consume.</commentary> </example> <example> Context: User references specific files to include user: "Use retro-notes.md and tool-audit.xlsx for context." assistant: "Reading retro-notes.md and tool-audit.xlsx. Extracting team structure, tooling landscape, and recurring problems..." <commentary>The ingester handles explicit file references by reading them directly and extracting the same structured fields.</commentary> </example>

Use this agent when the user wants to generate a complete playbook, or when the /playbook:forge command is invoked. This is the master orchestrator that coordinates all other playbook-forge agents through the full pipeline. <example> Context: User invokes the forge command with a topic user: "/playbook:forge GenAI adoption for QA teams" assistant: "Starting the playbook forge pipeline for GenAI adoption in QA. Let me gather some context before we begin..." <commentary>The orchestrator recognizes a topic-based invocation and begins the INTAKE phase, then proceeds to CLARIFY by asking targeted questions.</commentary> </example> <example> Context: User provides source files and requests a playbook user: "I have meeting notes and process docs in ./discovery/. Build me a playbook from them." assistant: "I found 12 files in ./discovery/. Launching the context ingester to extract structured data before we proceed..." <commentary>The orchestrator detects source files, launches context-ingester first, then uses the extracted brief to skip redundant questions.</commentary> </example>

Use this agent when a playbook manifest is ready and needs to be assembled into the final branded HTML file. This agent is purely deterministic -- it does not generate content, only assembles it. <example> Context: Assembling a playbook from a completed manifest user: "Assemble the playbook from outputs/.playbook-manifest.json" assistant: "Reading manifest... 13 flows, 4 katas, 12 sections found. Attempting Node assembly script first..." <commentary>The assembler tries the Node script path first for speed and consistency, falling back to manual snippet assembly if Node is unavailable.</commentary> </example> <example> Context: Assembling a single section for preview user: "Build just the hero and problems sections so I can preview the style." assistant: "Assembling partial playbook with hero and problems sections only. Using brand tokens from Sofka DS v5.1..." <commentary>The assembler supports partial builds for quick previews during iteration.</commentary> </example>

Genera playbooks HTML branded con metodologia vivida (13 flujos AI-native, 4 katas, adopcion progresiva, VR-AID, semaforo de valor). Use when the user asks to "crear playbook", "generar playbook", "playbook HTML", "cartilla de adopcion", "playbook de workflows", "generar cartilla".