invos 1.0.0 → 2.0.1

package/kit/.agents/skills/apple/apple-reminders/SKILL.md

@@ -1,130 +0,0 @@
----
-name: apple-reminders
-description: "Apple Reminders via remindctl: add, list, complete."
-version: 1.0.0
-author: Hermes Agent
-license: MIT
-platforms: [macos]
-metadata:
-  hermes:
-    tags: [Reminders, tasks, todo, macOS, Apple]
-prerequisites:
-  commands: [remindctl]
----
-
-# Apple Reminders
-
-Use `remindctl` to manage Apple Reminders directly from the terminal. Tasks sync across all Apple devices via iCloud.
-
-## Prerequisites
-
-- **macOS** with Reminders.app
-- Install: `brew install steipete/tap/remindctl`
-- Grant Reminders permission when prompted
-- Check: `remindctl status` / Request: `remindctl authorize`
-
-## When to Use
-
-- User mentions "reminder" or "Reminders app"
-- Creating personal to-dos with due dates that sync to iOS
-- Managing Apple Reminders lists
-- User wants tasks to appear on their iPhone/iPad
-
-## When NOT to Use
-
-- Scheduling agent alerts → use the cronjob tool instead
-- Calendar events → use Apple Calendar or Google Calendar
-- Project task management → use GitHub Issues, Notion, etc.
-- If user says "remind me" but means an agent alert → clarify first
-
-## Quick Reference
-
-### View Reminders
-
-```bash
-remindctl                    # Today's reminders
-remindctl today              # Today
-remindctl tomorrow           # Tomorrow
-remindctl week               # This week
-remindctl overdue            # Past due
-remindctl all                # Everything
-remindctl 2026-01-04         # Specific date
-```
-
-### Manage Lists
-
-```bash
-remindctl list               # List all lists
-remindctl list Work          # Show specific list
-remindctl list Projects --create    # Create list
-remindctl list Work --delete        # Delete list
-```
-
-### Create Reminders
-
-```bash
-remindctl add "Buy milk"
-remindctl add --title "Call mom" --list Personal --due tomorrow
-remindctl add --title "Meeting prep" --due "2026-02-15 09:00"
-```
-
-### Due Time vs Alarm / Early Nudge
-
-`--due` and `--alarm` are different fields:
-
-- `--due` sets the reminder's due date/time.
-- `--alarm` sets the EventKit alarm/notification trigger. Timed due reminders may default to an alarm at the due time, but pass `--alarm` explicitly when the user asks for an earlier nudge.
-
-For a reminder due at 2:00 PM with a notification 30 minutes earlier:
-
-```bash
-remindctl add --title "Hairdresser" --due "2026-05-15 14:00" --alarm "2026-05-15 13:30"
-```
-
-To edit an existing reminder:
-
-```bash
-remindctl edit 87354 --due "2026-05-15 14:00" --alarm "2026-05-15 13:30"
-```
-
-The Reminders UI may show or group the item by the alarm time because that is when the notification fires. Verify with JSON instead of assuming the due time moved:
-
-```bash
-remindctl today --json
-```
-
-Expected shape:
-
-- `dueDate`: actual due time
-- `alarmDate`: notification / early nudge time
-
-Apple's public `EKReminder` docs list only reminder-specific properties. Alarm support comes from inherited `EKCalendarItem` behavior exposed by remindctl's `--alarm` flag.
-
-### Complete / Delete
-
-```bash
-remindctl complete 1 2 3          # Complete by ID
-remindctl delete 4A83 --force     # Delete by ID
-```
-
-### Output Formats
-
-```bash
-remindctl today --json       # JSON for scripting
-remindctl today --plain      # TSV format
-remindctl today --quiet      # Counts only
-```
-
-## Date Formats
-
-Accepted by `--due` and date filters:
-- `today`, `tomorrow`, `yesterday`
-- `YYYY-MM-DD`
-- `YYYY-MM-DD HH:mm`
-- ISO 8601 (`2026-01-04T12:34:56Z`)
-
-## Rules
-
-1. When user says "remind me", clarify: Apple Reminders (syncs to phone) vs agent cronjob alert
-2. Always confirm reminder content and due date before creating
-3. Use `--json` for programmatic parsing

package/kit/.agents/skills/audit/SKILL.md

@@ -1,177 +0,0 @@
----
-name: audit
-description: Use for AIOS audit — scores project against Four Cs, surfaces leverage-weighted top-3 gaps. Read-only. Run on Day 7, then weekly. Watch score climb.
----
-
-## What this skill does
-
-Runs the **Four Cs Audit** on the current project. Reads (never writes) the project's operating manual, memory, skills, agents, connections, decisions, and references. Scores each of the Four Cs out of 25. Surfaces strengths and the top 3 leverage-weighted gaps with concrete next-step commands.
-
-**Scope is structural — "is the AIOS built right?"** It is NOT a capability planner. Capability gaps belong to `/level-up`.
-
-First run is the baseline. Re-run weekly to watch the score climb.
-
-## Today's context
-
-- **Date:** `date +%Y-%m-%d`
-- **Project root:** the current working directory
-
-## The Four Cs (scored 25 each = 100 total)
-
-| Layer | Test |
-|---|---|
-| **Context** | Knows the business — identity, team, voice, decisions, references |
-| **Connections** | Reaches the user's stuff — MCPs, integrations, data sources |
-| **Capabilities** | Knows how to do work — skills + agents |
-| **Cadence** | Runs without being asked — schedules, hooks, recurring rituals |
-
-## Execution
-
-### Step 1: Discover the project shape
-
-The audit looks for **patterns and intent**, not exact paths. Use Glob and Read to check:
-
-**Operating manual:** `CLAUDE.md` (root), `CLAUDE.local.md` (gitignored).
-**Memory:** `memoria/` folder.
-**Skills:** `.agents/skills/*/SKILL.md` — count + frontmatter. Also check `.claude/skills/*/SKILL.md` (mirror surface).
-**Agents:** `.claude/agents/*.md` — count + frontmatter.
-**Connection mechanisms** (any of these = reachable):
-- MCPs: `opencode.json` (mcpServers key), `.claude/settings.local.json` (mcpServers key), `.mcp.json` (root or subproject)
-- API scripts: `scripts/*.py|.js|.ts` documented in CLAUDE.md
-- Export pipelines: `data/`, `imports/`, `saidas/` with refresh script + last-run timestamp
-- API keys + reference guide: `.env` entries + corresponding `references/{tool}-api.md`
-- Vercel functions: `api/` directory
-
-**Connections registry:** `connections.md` (root).
-**Reference guides:** `references/` folder.
-**Decisions:** `decisions/log.md` (root).
-**SOPs:** `docs/` folder.
-**Templates:** `conhecimento/templates/`, `templates/`.
-**Hooks / scheduled jobs:** `.claude/hooks/` for session hooks; `.github/workflows/` for CI schedules.
-
-Don't penalize for non-canonical names if equivalent intent is captured elsewhere.
-
-### Step 2: Score each C (25 points each)
-
-#### Context (25 pts)
-
-| Criterion | Points | How to detect |
-|---|---|---|
-| Operating manual exists and is substantive (>200 words) | 5 | Read CLAUDE.md, count words |
-| Identity / role / voice captured | 5 | CLAUDE.md mentions who the user is + role/mission, OR `memoria/perfil.md` exists |
-| Persistent memory exists with multiple entries | 5 | `memoria/` has >3 files |
-| Reference docs exist | 5 | `references/`, `docs/`, or `conhecimento/voice/` has >=1 file |
-| Decisions captured | 5 | `decisions/log.md` or equivalent has >=1 entry |
-
-#### Connections (25 pts) — domain-aware, mechanism-agnostic
-
-A "reachable" connection counts via ANY mechanism: MCP, script, export pipeline, or `.env` key + `references/{tool}-api.md`.
-
-**The 7 Tier-1 Universal Data Domains:**
-
-| # | Domain | Examples |
-|---|---|---|
-| 1 | Revenue / Financials | Stripe, Skool, GoHighLevel, QuickBooks, Looker |
-| 2 | Customer interactions | HubSpot, Salesforce, Gmail-as-CRM, Skool DMs |
-| 3 | Calendar | Google Cal, Outlook, Calendly |
-| 4 | Communication | Gmail, Outlook, Slack, Teams, Telegram, Discord |
-| 5 | Project / task tracking | ClickUp, Asana, Linear, Notion DB, GitHub Issues |
-| 6 | Meeting intelligence | Granola, Otter, Fireflies, Gong, Zoom |
-| 7 | Knowledge / files | Notion, Drive, Dropbox, Confluence, GitHub repo |
-
-**Tier-2 (bonus):** AI service API keys, decisions/history, content/publishing.
-
-| Criterion | Points | How to detect |
-|---|---|---|
-| Tier-1 domain coverage | 10 | 1.4 pts per tier-1 domain reachable. Round to nearest 0.5. Cap 10. |
-| Reference guide presence | 5 | -1 per connected tool with no `references/{tool}-api.md`. Floor 0. |
-| Auth / pipeline freshness | 5 | -1 per connection in needs-auth/expired state, or script with no run within 30 days. Floor 0. |
-| Documentation in `connections.md` | 3 | 0 if missing; 1 sparse; 2 most; 3 covers all reachable. |
-| Read-AND-write balance | 2 | At least one connection can WRITE. 0 if all read-only. |
-
-#### Capabilities (25 pts)
-
-| Criterion | Points | How to detect |
-|---|---|---|
-| 3+ skills installed | 10 | Count `.agents/skills/*/SKILL.md` >= 3 |
-| 1+ user-built skill | 10 | Skill names not in canonical list: `onboard`, `audit`, `level-up`, plus any shipped Anthropic skills |
-| 1+ agent defined | 5 | Count `.claude/agents/*.md` >= 1 |
-
-#### Cadence (25 pts)
-
-| Criterion | Points | How to detect |
-|---|---|---|
-| 1+ recurring/scheduled trigger | 10 | `.github/workflows/*.yml` with cron/schedule, OR hook scripts in `.claude/hooks/` |
-| Recent activity / usage signal | 10 | Skills modified within 30 days, OR `decisions/log.md` has entry within 30 days |
-| Templates folder populated | 5 | `templates/` or `.claude/templates/` has >=1 file |
-
-### Step 3: Identify top 3 gaps by leverage
-
-For each criterion that lost points: leverage = (points lost) × (impact multiplier).
-
-**Impact multipliers:**
-- 0 tier-1 domains reachable: **4x** (AIOS is blind to the business)
-- Operating manual missing or thin: **3x** (foundation)
-- <=2 tier-1 domains reachable: **3x** (Connections is the gateway to live data)
-- 0 skills: **2x** (no Capabilities = no AIOS)
-- No recurring trigger: **2x** (no Cadence = no autonomy)
-- All connections read-only: **2x** (viewer, not an OS)
-- 0 reference guides for connected tools: **1.5x**
-- No decisions log: **1.5x**
-- All others: **1x**
-
-Sort gaps by leverage descending. Take top 3. For each, write a one-line concrete next step.
-
-### Step 4: Output the report
-
-Print directly in chat (Markdown). Format:
-
-```
-# AIOS Audit — {date}
-**Score: {total}/100** ({stage})
-
-Stage thresholds:
-- 0-39 → Stage 0: Foundation
-- 40-69 → Stage 1: Built
-- 70-89 → Stage 2: Compounding
-- 90-100 → Stage 3: Autonomous
-
-## Scoreboard
-
-Context        {bar}  {n}/25  {label}
-Connections    {bar}  {n}/25  {label}
-Capabilities   {bar}  {n}/25  {label}
-Cadence        {bar}  {n}/25  {label}
-
-(bar = ## per 5pts; label = "Strong" >=20, "Solid" 15-19, "Thin" 8-14, "Missing" <8)
-
-## Strengths
-- {1-3 short bullets from highest-scoring criteria}
-
-## Top 3 Gaps (ranked by leverage)
-1. **{gap name}** (-{n}pts x {multiplier})
-   -> {concrete next-step}
-2. **{gap name}** (-{n}pts x {multiplier})
-   -> {concrete next-step}
-3. **{gap name}** (-{n}pts x {multiplier})
-   -> {concrete next-step}
-
-## Suggested next: {single most leveraged action}
-
----
-Structural gaps only. To explore CAPABILITY gaps, run /level-up after this audit.
-```
-
-### Step 5: Offer to save the report
-
-After printing, ask: "Save this audit to `audits/audit-{date}.md` so you can track score over time?" If yes, write it. This is the only writable side effect.
-
-## Notes
-
-- **Read-only by default.** Never modify CLAUDE.md, memory, skills, or any project files. Only optional write is the audit report.
-- **Be flexible about file names.** Don't penalize for using non-canonical names if intent is captured.
-- **Be honest, not generous.** A 95/100 is rare. Most setups land 40-70.
-- **Speed matters.** Report in under 60 seconds wall-clock.
-- **Cadence detection is fuzzy.** Infer from workflows and hook scripts if cron data isn't cleanly available.
-
-> *Adapted from The Three Ms of AI™ © 2026 Nate Herk.*

package/kit/.agents/skills/design-marca/SKILL.md

@@ -1,92 +0,0 @@
----
-name: design-marca
-description: >-
-  Extrai ou atualiza a identidade visual do negócio em marca/marca.md (cores,
-  tipografia, assets). Use quando: "extrai a marca do meu site", "design system",
-  "/design", "atualiza cores da marca", "pega o branding do site X".
----
-
-# Design → marca/ (INVOS)
-
-Adaptação da ideia **design-md** pro INVOS: o destino **não** é `DESIGN.md` na raiz.  
-Destino canônico = **`marca/marca.md`** (+ `marca/assets/` se houver logo).
-
-Objetivo: proposta HTML e qualquer saída usarem **as cores e a voz reais** do prestador.
-
-## Gatilhos
-
-- “Extrai a marca do site …”, “/design …”, “atualiza brand”, “minhas cores”
-- Onboard ofereceu preencher marca e o user mandou URL
-- Proposta pediu brand e `marca/marca.md` ainda é template
-
-## Fontes de dados (ordem)
-
-1. **URL do site do usuário** (preferido se tiver)  
-2. Respostas manuais (nome, cor favorita, tom)  
-3. Já existente em `marca/marca.md` / `memoria/empresa.md` (não apagar o que for bom)
-
-## Extração de site (se houver URL)
-
-Normalizar URL (`https://` se faltar).
-
-Tentar, nesta ordem:
-
-1. MCP/CLI Hyperbrowser se existir no ambiente:  
-   `hx web fetch <url> --format branding`  
-   ou tool `scrape_webpage` com branding  
-2. Se **não** houver Hyperbrowser / API key:  
-   - abrir a página com ferramenta de browse disponível, **ou**  
-   - perguntar 4 coisas manuais: cor principal (hex), cor de texto, fonte se souber, tom visual  
-   - **Não inventar** paleta “bonita” sem base
-
-**Regra:** só grave hex/fontes que foram **extraídos ou ditos pelo user**. Sem alucinar design system completo.
-
-## O que gravar
-
-### Sempre atualizar `marca/marca.md`
-
-Preencher / mesclar seções:
-
-- Identidade (nome da marca, site, contato se souber)  
-- Cores (primária, fundo, texto — hex)  
-- Tipografia (se veio na extração)  
-- Voz: se o user já tem tom em `empresa.md`, **copie** pra cá (não deixe só em empresa)
-
-Mantenha o formato do template. Não reescreva o arquivo apagando campos úteis sem necessidade.
-
-### Logo
-
-- Se o user **anexar** logo: salvar em `marca/assets/logo.png` ou `logo.svg` e marcar **Logo presente? sim**  
-- Se só URL de logo clara e download for possível/seguro: salvar em `assets/`  
-- Senão: deixar “não” e seguir só com cor + nome
-
-### Espelhar o mínimo em `memoria/empresa.md`
-
-Atualize a seção **Brand** (ou o ponteiro) pra não divergir:
-
-- Marca, Cor, Site, Contato = mesmos valores de `marca/marca.md`
-
-`marca/marca.md` continua sendo a **fonte**; empresa só resume.
-
-## Output no chat
-
-```
-✅ Marca atualizada em marca/marca.md
-- Primária: #…
-- Fundo / texto: …
-- Logo: sim (path) | não
-Próximo: gerar proposta ou post já usa essas cores.
-```
-
-## O que NÃO fazer
-
-- Salvar `DESIGN.md` na raiz do INVOS (confunde com o core)  
-- Inventar cores/fontes “de gosto”  
-- Misturar brand do **cliente** com a marca do prestador  
-- Depender de API paga sem fallback manual  
-
-## Relação com squad `brand`
-
-Squad brand = consultoria profunda (posicionamento, brand book).  
-Esta skill = **operacional**: 5 minutos, arquivo que a proposta lê.  
-Se o squad gerar brand book, **resuma** cores/voz/logo path em `marca/marca.md`.

package/kit/.agents/skills/instagram-carrossel/SKILL.md

@@ -1,79 +0,0 @@
----
-name: instagram-carrossel
-description: >-
-  Instagram no INVOS: carrossel (default MVP 1–2 slides HTML + caption + handoff),
-  opcional render PNG, opcional publish Graph API. Use quando: carrossel, slides IG,
-  fábrica de carrossel, pack carrossel, render PNG dos slides, publica no Instagram
-  (Business + .env). NÃO use para só legenda texto (social-content).
----
-
-# Instagram carrossel (pacote único)
-
-Uma skill de entrada. Dependências antigas (`image-creator`, `image-fetcher`, `instagram-publisher`) vivem em **`references/`** — o agente carrega o subdoc **só quando o passo exigir**.
-
-```text
-User: carrossel / slides
-        → este SKILL (MVP ou full)
-User: gera PNG / render
-        → references/render-png.md
-User: busca imagem / screenshot pra slide
-        → references/fetch-assets.md
-User: publica no IG (API)
-        → references/publish.md + scripts/publish.js
-```
-
----
-
-## Default: modo MVP (dia a dia)
-
-Prestador em 5–15 min — **não** rode 13 steps do squad.
-
-1. Leia `marca/marca.md` + `memoria/empresa.md` (+ prova em `clientes/` se houver).  
-2. Pasta `conteudo/carrosseis/YYYY-MM-DD-kebab/`.  
-3. Entregue:
-   - `caption.txt`
-   - `slide-01.html` (+ opcional `slide-02.html`) — 1080×1350, cores da marca  
-   - `README.md` (abrir browser / print / ⚠️ postar no app)  
-4. `conteudo/_fila.md` atualizado.  
-5. Chat: paths + `⚠️ Sua vez: subir no Instagram`.
-
-**PNG:** só se user pedir **e** Chrome/Python ok → siga `references/render-png.md`.  
-Senão: HTML + handoff. **Não finja** pipeline completo.
-
-## Fábrica completa (opcional)
-
-Só se user disser: **“fábrica completa”**, **“modo squad”**, **“pipeline full”**, **“com research e ângulos”**.
-
-1. `.agents/squads/instagram-carrossel/RUN.md`  
-2. `squad.yaml` + `pipeline/`  
-3. Preferir modo turbo do RUN.
-
-## Publicar no Instagram (opcional)
-
-Só se user disser **“publica” / “pode postar”** **e** `.env` com tokens Business.
-
-1. Leia `references/publish.md`  
-2. Script: `node --env-file=.env .agents/skills/instagram-carrossel/scripts/publish.js …`  
-3. Sem env / falha → handoff humano. **Nunca** invente permalink.
-
-## Atalhos
-
-| User diz | O que rodar |
-|----------|-------------|
-| “faz um carrossel” / tema | **MVP** deste arquivo |
-| “gera PNG” / “exporta slides” | MVP (se falta HTML) + `references/render-png.md` |
-| “busca imagem pro slide” | `references/fetch-assets.md` |
-| “publica no IG” | pack pronto + `references/publish.md` ou ⚠️ |
-| “só legenda” | **Não** esta skill → `social-content` |
-
-## Relação
-
-| Skill / path | Quando |
-|--------------|--------|
-| `social-content` | Texto/legenda sem slides |
-| `design-marca` | Marca vazia |
-| Squad `instagram-carrossel` | Só no modo full |
-
-## Dono
-
-Mínimo: subir pack no app. API e PNG são extras.

package/kit/.agents/skills/instagram-carrossel/references/fetch-assets.md

@@ -1,77 +0,0 @@
-# Fetch assets (subdoc de instagram-carrossel)
-
-**Não é skill separada.** Carregar só se precisar stock/screenshot/arquivo do user nos slides.
-
-## When to use
-
-Assets visuais: busca web, screenshot (Playwright **ou** Chrome se existir), arquivos do user.  
-**Opcional** no carrossel — a maioria das runs só precisa de HTML tipográfico + `marca/`.
-
-## When to use (detail)
-
-Three modes: web image search, live website screenshots, user-provided files. Save under squad `output/` or `conteudo/carrosseis/$SLUG/assets/`.
-
-## Instructions
-
-### Capabilities
-
-1. **Web Image Search** -- Use the native web_search tool to find images by keyword. Evaluate results and download the best match.
-
-2. **Live Screenshot** -- Use Playwright MCP to navigate to a URL, set viewport dimensions, and capture a screenshot.
-
-3. **Asset Organization** -- Save all acquired assets with descriptive filenames in the squad's reference/ or output/ folder.
-
-### Screenshot Modes
-
-- **viewport** -- Capture only the visible viewport area (default)
-- **full_page** -- Capture the entire scrollable page
-- **selector** -- Capture a specific CSS selector element
-
-### Screenshot Workflow
-
-When taking a screenshot:
-1. Navigate to the URL with `browser_navigate`
-2. Set viewport: `browser_resize` with width/height for target format
-   - Instagram post: 1080x1080
-   - Instagram carousel: 1080x1440
-   - Story/Reel: 1080x1920
-   - Generic: 1280x720
-3. Wait for page load (`browser_wait_for` if needed)
-4. Capture: `browser_take_screenshot`
-5. Save to reference folder with descriptive filename
-
-### Asset Metadata
-
-After acquiring each asset, document in your output:
-- `path`: local file path
-- `width/height`: image dimensions
-- `source_type`: "web_search" | "screenshot" | "user_provided"
-- `original_url`: source URL (if applicable)
-
-### Cache Policy
-
-Before fetching an asset:
-- Check if the reference folder already has a matching file
-- Use deterministic filenames based on source (e.g., URL slug + viewport)
-- Reuse existing assets to avoid redundant fetches
-
-### Safety
-
-- Timeout: max 30s per screenshot, skip and warn if exceeded
-- Maximum screenshot dimensions: 1920x1920px
-- Block `file://` protocol URLs
-- Block localhost and private IP ranges (127.0.0.1, 10.x, 192.168.x)
-
-### Best Practices
-
-- Prefer screenshots over web search for product/tool pages (images are often outdated)
-- Save with descriptive names: `gemini-benchmark-chart.png` not `image1.png`
-- Normalize URLs before caching (strip tracking params)
-- Document all acquired assets with metadata for downstream tools
-
-## Available operations
-
-- **Web Image Search** -- Find and download images by keyword from the web
-- **Live Screenshot** -- Capture viewport, full-page, or element screenshots of any URL
-- **Asset Organization** -- Save and catalog assets with descriptive filenames and metadata
-- **Cache Check** -- Detect and reuse previously fetched assets to avoid redundant downloads

package/kit/.agents/skills/instagram-carrossel/references/publish.md

@@ -1,48 +0,0 @@
-# Publish Instagram (subdoc de instagram-carrossel)
-
-**Não é skill separada.** Carregar só se user disser “publica/pode postar” e houver `.env`.
-
-## Default no INVOS (sem erro se faltar API)
-
-**Não é obrigatório.** O skill `instagram-carrossel` termina com pack em `conteudo/carrosseis/` +  
-
-```text
-⚠️ Sua vez: subir PNGs no Instagram + colar caption.txt
-```
-
-Só use publish automático se:
-
-1. User disse **“publica” / “pode postar”**  
-2. `.env` tem as variáveis (nunca commitar)  
-3. Conta Instagram **Business** + Página Facebook  
-
-## Env (local only)
-
-```bash
-# .env (gitignore)
-INSTAGRAM_ACCESS_TOKEN=
-INSTAGRAM_USER_ID=
-IMGBB_API_KEY=
-```
-
-## Script
-
-```bash
-node --env-file=.env .agents/skills/instagram-carrossel/scripts/publish.js \
-  --images "path/slide-01.png,path/slide-02.png" \
-  --caption "legenda..."
-# dry-run:
-#   ... --dry-run
-```
-
-Imagens: JPEG/PNG conforme o script; preferir 2–10 slides. Caption max 2200 chars.
-
-## Setup token (resumo)
-
-Ver Meta Graph API Explorer: permissões `instagram_content_publish`, `instagram_basic`, `pages_read_engagement`.  
-User ID: Page → `instagram_business_account.id`.  
-Token longa duração (~60 dias).
-
-## Se faltar chave / script falhar
-
-Reportar erro real → manter handoff humano. **Nunca** inventar permalink de post.