Construir verificacion cruzada declarado-vs-calculado con mismatch flag tipado y escalada a humano; recomputo independiente, epsilon justificado, nunca corregir numeros en silencio.

Router for application security playbooks: auth, RBAC, input sanitization, headers/CORS, rate limiting, audits, security testing. Resolve one topic, load one reference. Topics: architecture, audit-security, auth-architecture, cors-configuration, dual-layer-verification, http-headers, input-sanitization, rate-limiting, rbac-patterns, testing.

Router for sales and business development in an ES/EN consulting context — prospecting, outreach, dossiers, proposals, pitches, and collateral. Topics: b2b-outreach, client-dossier, client-prospecting, executive-pitch, lead-generation, proposal-writing, sales-collateral.

Design and review safe, portable, dry-run-first Bash scripts for local agentic workflows; gate writes, destructive commands, secrets, and offline validation.

Route agentic work across Claude, Codex, Gemini, Antigravity, VS Code, and local adapters, picking the lowest-permission runtime that has evidence for the task and labeling every unverified capability.

Audita un texto, etiqueta cada afirmacion no reproducible con un tag de veracidad y propone el siguiente paso concreto de verificacion.

Cadencia QBR (P13): audita el trimestre cerrado contra OKRs y planifica el proximo con evidencia trazable.

Disenar pipelines de extraccion/sintesis con provenance tipada: invariante no hay claim sin source, conflictos marcados y escalados a humano, nunca promediados ni resueltos en silencio.

Turn vague intentions into durable, eval-ready prompts, meta-prompts, acceptance criteria, and prompt systems with anti-drift and safety gates.

Descomponer tareas grandes en pase local tipado y pase de integracion sobre resumenes, con schemas de transicion entre pases.

Scaffold a Jarvis OS project P-NNN-slug under 02_Proyectos/ with compliant CLAUDE.md, MEMORY.md and TAREAS.md (P07); missing-only, never clobbers local edits.

Product and business analytics router: event instrumentation, KPIs, A/B tests, cohorts, funnels, real-time pipelines, and data visualization. Topics: ab-testing, analytics-events, cohort-analysis, data-visualization, funnel-analytics, kpi-framework, metrics-instrumentation, real-time-analytics.

Forja contenido brand-compliant MetodologIA desde lenguaje natural. Combina Brand Voice v4.0 (Minto-First + lista roja con prohibición de "gratis"/ "gratuito") con estética Neo-Swiss Clean and Soft Explainer (paleta exclusiva navy #122562 / gold #FFD700 / blue #137DC5 / dark #1F2833 / lilac #BBA0CC / gray #808080 + Poppins/Montserrat/Trebuchet). Genera Markdown, HTML self-contained, slide decks, posts LinkedIn, emails, playbooks. Activa para frases como "post MetodologIA", "playbook MetodologIA", "audita esta pieza con voz MetodologIA", "genera HTML brand-ready MetodologIA", "crea contenido con voz Minto", "reescribe esto con voz MetodologIA". Use proactively cuando el usuario pida contenido educativo o de marca MetodologIA.

Use proactively when the user asks to "aprender un tema", "ponerme al día", "hacer deep research", "preparar una certificación, entrevista o QBR", "configurar NotebookLM", "comprobar una investigación" or "auditar qué conocimientos siguen vigentes". Routes evidence-based learning through Aprender, Aprehender, (R)Evolucionar or Auditoría without promising mastery, treating model agreement as evidence, or persisting user state by default.

Handoff a humano con payload tipado autocontenido (customer_id, issue_summary, actions_taken); end-state del bucle, no pausa.

This skill should be used when the user asks to operate Claude Code /goal, write a goal completion condition, compare /goal with /loop, Stop hooks, or auto mode, diagnose a stuck goal loop, set stop conditions, resume or clear an active goal, or define safe continuation gates for autonomous Claude Code sessions.

This skill should be used when the user asks to automate Claude Code headlessly, run claude -p in CI or scripts, parse json or stream-json output, use --bare, configure non-interactive permissions, compare CLI vs TypeScript or Python Agent SDK usage, wire MCP or hooks for automation, resume sessions programmatically, or harden reproducible Claude Code automation.

This skill should be used when the user asks to scaffold or verify a Claude Agent SDK app in Python or TypeScript — routing to the agent-sdk-dev plugin under toolkit governance.

This skill should be used when the user asks to audit, improve, or update CLAUDE.md files with session learnings — routing to the claude-md-management plugin under toolkit governance.

This skill should be used when the user asks to commit, push and open a PR, or clean gone git branches — routing to the commit-commands plugin under toolkit governance.