wizz-method 2.17.2 → 2.19.0

package/src/modules/lowticket/skills/wizz-lowticket-funil/SKILL.md

@@ -17,7 +17,7 @@ Every offer in the ladder exists because it answers that question for the step b
 
 ## Source of truth
 
-Read the Cérebro state first (via the `cerebro` skill). Then, if `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md` and the shard `07-funil-upsell.md`, especially §7.4 (the conversion and ticket cascade) and §7.5-7.9 (pain per step, tone per page, copy rules, the technical block for post-checkout pages and the pre-build checklist). Its sibling shard `07a-copy-humanizacao.md` carries the humanization checklist used before delivering any page. Where a generic skill contradicts that shard, the shard wins and you say so.
+Read the Cérebro state first (via the `wizz-cerebro` skill). Then, if `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md` and the shard `07-funil-upsell.md`, especially §7.4 (the conversion and ticket cascade) and §7.5-7.9 (pain per step, tone per page, copy rules, the technical block for post-checkout pages and the pre-build checklist). Its sibling shard `07a-copy-humanizacao.md` carries the humanization checklist used before delivering any page. Where a generic skill contradicts that shard, the shard wins and you say so.
 
 This skill has **no memory of its own**. What gets decided goes to the Cérebro; the artifact goes to a file.
 

package/src/modules/lowticket/skills/wizz-lowticket-guia/SKILL.md

@@ -11,7 +11,7 @@ The other low ticket skills are sharp and assume the user knows which one to cal
 
 ## Step 0: What is already known
 
-1. Read the project's Cérebro state (via the `cerebro` skill). If it already says which offer is live and at which step, skip the questions.
+1. Read the project's Cérebro state (via the `wizz-cerebro` skill). If it already says which offer is live and at which step, skip the questions.
 2. Read `wizz-lowticket-metodologia/knowledge/INDEX.md` and `guardrails.md` (the always-load of the base). Read `01-principios.md` too: its gates decide which steps cannot be skipped.
 
 ## Step 1: Place the user, in one message

package/src/modules/lowticket/skills/wizz-lowticket-metodologia/SKILL.md

@@ -31,7 +31,7 @@ The one thing that outranks a shard is `knowledge/guardrails.md`. Eight tactics
 
 ### Step 0: Read the state, then the index
 
-Read the project's Cérebro state block first (via the `cerebro` skill: current project state plus at most 3 decisions on the topic). This skill has **no memory of its own** and never writes one. What is decided lives in the Cérebro; what is known lives in the shards.
+Read the project's Cérebro state block first (via the `wizz-cerebro` skill: current project state plus at most 3 decisions on the topic). This skill has **no memory of its own** and never writes one. What is decided lives in the Cérebro; what is known lives in the shards.
 
 Then read `knowledge/INDEX.md` and `knowledge/guardrails.md`. That is the whole default load, around 800 tokens.
 

package/src/modules/lowticket/skills/wizz-lowticket-minerador/SKILL.md

@@ -15,7 +15,7 @@ Two kinds of person ask this. An operator who already runs traffic, reads metric
 
 ## Source of truth
 
-Read the Cérebro state first (via the `cerebro` skill). If `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md`, then:
+Read the Cérebro state first (via the `wizz-cerebro` skill). If `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md`, then:
 
 - **`12-mineracao-ofertas.md`**: the price trick, the Ad Library filter checklist and keyword taxonomy, the 0-10 rubric with a cut at 6, the three safety rules
 - **`12a-mineracao-veredito.md`**: §12.5, the scoring ruler and what disqualifies

package/src/modules/lowticket/skills/wizz-lowticket-pagina/SKILL.md

@@ -30,7 +30,7 @@ Speed fixes the leak between the click and the page. It does not fix a bad offer
 
 ## Source of truth
 
-Read the Cérebro state first (via the `cerebro` skill). If `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md` and shard `06-pagina-vendas.md`, especially §6.3-6.4. Page structure and copy blocks for a pure low-ticket page live in `12c-anatomia-low-ticket.md` (the 11-block anatomy, not `12b` anymore); VSL player behavior (autoplay, progress bar, weight) is `06b-player-vsl.md`; conversion-side reasoning is `wizz-page-cro`. Where a generic skill contradicts the shard, the shard wins and you say so.
+Read the Cérebro state first (via the `wizz-cerebro` skill). If `wizz-lowticket-metodologia` is installed, read `knowledge/INDEX.md` + `knowledge/guardrails.md` and shard `06-pagina-vendas.md`, especially §6.3-6.4. Page structure and copy blocks for a pure low-ticket page live in `12c-anatomia-low-ticket.md` (the 11-block anatomy, not `12b` anymore); VSL player behavior (autoplay, progress bar, weight) is `06b-player-vsl.md`; conversion-side reasoning is `wizz-page-cro`. Where a generic skill contradicts the shard, the shard wins and you say so.
 
 **Open conflict, do not resolve it alone.** §6.3-6.4 (this skill's rule) fires the pixel direct, first in `<head>`, no tag manager. `17-rastreamento.md` installs that same pixel *inside* GTM, as part of the full GA4 + Meta + server-container stack. This is ponto 38 of `pontos-em-aberto.md` in `wizz-lowticket-metodologia`, and it is unresolved. If the page already runs tracking through GTM, do not rip the pixel out of the container to chase the speed target: ask which rule wins before touching it.
 

package/src/modules/lowticket/skills/wizz-lowticket-trafego/SKILL.md

@@ -18,7 +18,7 @@ The numbers live in `wizz-lowticket-metodologia`, not here. On activation, read
 
 Where a generic paid-media skill (`wizz-paid-ads`, `wizz-analytics-tracking`) contradicts those shards, the shard wins and you say so. If `wizz-lowticket-metodologia` is not installed, the rules in `references/reguas.md` here are the fallback, and you say that too.
 
-Read the Cérebro state before answering (via the `cerebro` skill): the account's breakeven, ticket and current bid cap are state, not knowledge. This skill has **no memory of its own** and writes none.
+Read the Cérebro state before answering (via the `wizz-cerebro` skill): the account's breakeven, ticket and current bid cap are state, not knowledge. This skill has **no memory of its own** and writes none.
 
 ## If the user is new to this
 

package/src/modules/wizz/_shared/cerebro.md

@@ -8,7 +8,7 @@ There are two layers, and each fact lives in exactly one: the **vault** (decisio
 
 Load context **cheaply**:
 
-1. Run `/cerebro ver` (or a light `grep` on the vault) to pull the project's current state and recent decisions.
+1. Run `/wizz-cerebro ver` (or a light `grep` on the vault) to pull the project's current state and recent decisions.
 2. Don't read the whole vault. Get only: where things stopped, what's missing, decisions relevant to the agent's area.
 3. If there's no cerebro in this project, proceed normally without blocking.
 
@@ -17,10 +17,10 @@ Load context **cheaply**:
 - If something worth remembering happened (architecture/product decision, learning, resolved gotcha, correction from the user), add to the closing:
   `💾 Quer que eu salve isso no cerebro?`
 - If the user confirms, use the right cerebro command:
-  - `/cerebro decisao` — architecture/product decision
-  - `/cerebro salvar` — session state
-  - `/cerebro dump` — quick capture
-  - `/cerebro memoria` — gotcha or preference (auto-memory), when the agent isn't Claude Code
+  - `/wizz-cerebro decisao` — architecture/product decision
+  - `/wizz-cerebro salvar` — session state
+  - `/wizz-cerebro dump` — quick capture
+  - `/wizz-cerebro memoria` — gotcha or preference (auto-memory), when the agent isn't Claude Code
 
 ## Duty to remember (non-negotiable)
 
@@ -30,8 +30,8 @@ After a non-obvious bug is resolved, an important decision, a stack gotcha, or a
 
 | type of fact | goes into |
 |---|---|
-| architecture or product decision | Cerebro `_decisions/` (`/cerebro decisao`) |
-| project state (where it stopped, what's missing) | Cerebro `projetos/<name>.md`, block at the top (`/cerebro salvar`) |
+| architecture or product decision | Cerebro `_decisions/` (`/wizz-cerebro decisao`) |
+| project state (where it stopped, what's missing) | Cerebro `projetos/<name>.md`, block at the top (`/wizz-cerebro salvar`) |
 | session narrative (what happened that day) | Cerebro `projetos/<name>.md`: `Estado` block at the top (overwritten) + 1 line in the `Sessões` table |
 | stack gotcha, non-obvious bug, environment quirk | agent auto-memory (`memory/` of the project; outside Claude Code, via `/memoria`) |
 | work preference, correction received from the user | agent auto-memory (same) |

package/src/modules/wizz/_shared/communication-rules.md

@@ -15,6 +15,6 @@
 
 WIZZ COMMUNICATION: don't narrate the step-by-step while working, it wastes tokens and clutters the conversation. Work directly. PAUSE and ask before continuing only when a decision belongs to the user, or something carries irreversible risk (deleting, publishing, overwriting, spending money). Always close with a short summary, in bullet points, in simple, easy-to-understand language. Don't use em-dashes in text a third party reads (public page, copy, UI, email): prefer a period, comma, colon, or parentheses. Code comments, commits, internal docs, and conversation with the user are exempt from this check: don't spend tokens reviewing or swapping em-dashes there. Concise reply by default, in easy `{communication_language}` (the language chosen at install): expand only when asked. Don't repeat what the user already knows, don't re-explain what was already said in the session, and don't list options you won't follow (recommend and act).
 
-WIZZ CLOSING: finish every task in this format (labels shown in Portuguese; write them in `{communication_language}`, keeping the emojis): '✅ O que fiz' (1-2 simple sentences) + '➡️ Próximo passo' (recommended agent or action) + '🎯 Comando: /wizz:<something>'. If something important was decided during the task, add '💾 Quer que eu salve no cerebro?'.
+WIZZ CLOSING: finish every task in this format (labels shown in Portuguese; write them in `{communication_language}`, keeping the emojis): '✅ O que fiz' (1-2 simple sentences) + '➡️ Próximo passo' (recommended agent or action) + '🎯 Comando: /<exact-skill-name>' (lowercase with hyphen, e.g. `/wizz-maestro`; never `/wizz:maestro`). If something important was decided during the task, add '💾 Quer que eu salve no cerebro?'.
 
 RTK: RTK (Rust Token Killer) already rewrites shell commands automatically via hook. No need to invoke anything separately. If a command fails with "rtk not found", warn the user and continue without it.

package/src/modules/wizz/_shared/despacho-barato.md

@@ -6,4 +6,4 @@
 # once. Don't copy this text into another toml: reference it via `include`.
 # ---
 
-Mechanical work (scanning, lookup, repetitive diff, an already-specified change): don't run it in the session when it's large; dispatch wizz-exec-mechanical/wizz-exec-standard subagents, partitioned one per file/module, up to 4 concurrent. The session's model is the boss: it does common work itself, dispatches, and decides with the user. Hard work escalates to wizz-exec-strong when the session's own tier isn't enough (same error twice, cause unclear, or architecture/schema/auth/payments/3+ modules); strong never dispatches, it returns ESCALAR to the session, which then dispatches wizz-exec-ultra with strong's findings. Diff review always goes to wizz-exec-review (pinned to the Strong tier, never inherits the session). Tell the user the tier and quantity of each dispatch. Full ladder in `_shared/model-ladder.md` and the rule in `_shared/token-economy.md` of the wizz module.
+Mechanical work (scanning, lookup, repetitive diff, an already-specified change): don't run it in the session when it's large; dispatch wizz-exec-mechanical/wizz-exec-standard subagents, partitioned one per file/module, up to 4 concurrent. The session's model is the boss: it does common work itself, dispatches, and decides with the user. Hard work escalates to wizz-exec-strong when the session's own tier isn't enough (same error twice, cause unclear, or architecture/schema/auth/payments/3+ modules); strong never dispatches, it returns ESCALAR to the session, which then dispatches wizz-exec-ultra with strong's findings. Before any dispatch, measure the size, not just the difficulty: more than one layer (database, backend/worker, app/UI, external automation) or ~8 files means one executor per layer, each brief with max_calls at about 2/3 of the tier's turn cap; a cross-layer story never goes whole to one executor, and a partial return gets split, not "continue". Diff review always goes to wizz-exec-review (pinned to the Strong tier, never inherits the session). Tell the user the tier and quantity of each dispatch. Full ladder in `_shared/model-ladder.md` and the rule in `_shared/token-economy.md` of the wizz module.

package/src/modules/wizz/_shared/encerramento.md

@@ -13,12 +13,13 @@ Every time you finish a task or explain a service, close **exactly** in this for
 ➡️ Próximo passo
 <which agent to call OR what to do now>
 
-🎯 Comando: /wizz:<agent-or-action>
+🎯 Comando: /<skill-name>
 ```
 
 ## Rules
 
 - **Summarize in plain human language.** No "I implemented the refactor of module X." Prefer "I made the site faster and organized the code."
+- **The 🎯 command is the exact skill name, lowercase, with a hyphen:** `/wizz-maestro`, `/wizz-designer`, `/wizz-agent-dev`, `/wizz-cerebro`. Never `/wizz:maestro`, `/Wizz:maestro` or any `namespace:name` form: that command doesn't exist and the user can't run it. If the next step isn't a skill, write the plain action instead of inventing a command.
 - **Always point to ONE clear next step.** If there are options, state the recommended one first: "I recommend calling wizz-agent-dev to build it. For a quick tweak, use wizz-quick-dev. To see the visual first, call wizz-designer."
 - **Automatic chaining.** If you're part of a sequence already announced (by the maestro or by the planning gate), trigger the next agent/skill automatically right after your ✅. Pause ONLY for a business decision that belongs to the user or an irreversible risk (deploy, delete, spending money). Outside an announced sequence, suggest the next step and wait.
 - **If the task opens up work from another area**, say which agent covers it (e.g., "this one is copy → wizz-copy").
@@ -27,7 +28,7 @@ Every time you finish a task or explain a service, close **exactly** in this for
 
 ## Feature closing checklist (when a feature delivery closes)
 
-When the task closes a feature's delivery (deploy or owner acceptance), run `node _wizz/hooks/wizz-feature.js verificar` BEFORE the block above. It checks: feature resolved, artifacts in their feature folder, spec with acceptance criteria, `## Implementação` filled, roteiro with Rollback when the status is "no-ar"; and returns two items to confirm: Cérebro updated (state + one Sessions line) and auto-memory holding only traps or preferences. **Don't close the sequence with a pending item**: resolve it, or list it under `⚠️ Não coberto` only when the owner explicitly accepted it. Then `wizz-feature.js encerrar`. Full duty in `_shared/organizacao.md`.
+When the task closes a feature's delivery (deploy or owner acceptance), run `node _wizz/hooks/wizz-feature.js verificar` BEFORE the block above. It checks: feature resolved, artifacts in their feature folder, spec with acceptance criteria, `## Implementação` filled, roteiro with Rollback when the status is "no-ar"; and returns two items to confirm: Cérebro updated (state + one Sessions line) and auto-memory holding only traps or preferences. **Don't close the sequence with a pending item**: resolve it, or list it under `⚠️ Não coberto` only when the owner explicitly accepted it. Then `wizz-feature.js encerrar --cerebro`: it closes the feature and writes the Cérebro itself (state + one Sessions line); if it reports the vault wasn't found, record it by hand. Full duty in `_shared/organizacao.md`.
 
 ## Real example
 
@@ -39,7 +40,7 @@ Montei o plano da sua landing em 4 blocos: topo, prova social, oferta e rodapé.
 Recomendo o wizz-designer pra desenhar o visual.
 Se preferir já partir pro código, chame o wizz-agent-dev.
 
-🎯 Comando: /wizz:designer
+🎯 Comando: /wizz-designer
 💾 Quer que eu salve a estrutura no cerebro?
 ```
 

package/src/modules/wizz/_shared/model-ladder.md

@@ -53,6 +53,16 @@ Review workflows (`wizz-code-review`, `wizz-quick-dev`) stay pinned to the Stron
 - Default cap: 4 concurrent, 10 per request. Going past 10 dispatches in one request, or past 4 concurrent, requires asking the user first. Never two on the same file.
 - "Large" mechanical work is more than 3 files, or a scan across a directory — it still doesn't run in the session. A small mechanical change can run in the session.
 
+## Size before dispatch
+
+The tier answers "how hard"; it does not answer "how big". Work that is easy per file but wide (several layers) fits the tier and still runs out of turns. Before dispatching, the session measures the size:
+
+1. **Count layers and files.** A layer is one of: database (migration, SQL, pgTAP), backend or worker, app or UI, external automation (n8n and similar), docs or script.
+2. **More than ~8 files OR more than one layer = split.** One executor per layer, with disjoint file sets. When one layer depends on another (database before app), dispatch them in sequence and pass the contract (function, table, and route names) in the next brief.
+3. **Budget in every brief.** A split dispatch always carries `max_calls` below the tier's turn cap: about 2/3 of it (35 for mechanical, 55 for standard, 80 for strong, 100 for ultra). The brief tells the executor to stop and report what is left when it gets close, instead of dying at the limit.
+4. **A whole story never goes to a single executor** when it crosses layers. The session (the boss) joins the parts, runs the suite, and makes the single commit.
+5. **A resume is not a plan.** If an executor comes back partial because of a limit, the session splits what is left into new briefs instead of sending "continue" to the same agent.
+
 ## Transparency
 
 Every dispatch states the tier and quantity in the text to the user ("1 × strong", "3 × mechanical"). Never just "executors".

package/src/modules/wizz/_shared/organizacao.md

@@ -7,6 +7,8 @@ Every step below is a **command**, not a memory exercise. The commands live in t
 - `F` = `node _wizz/hooks/wizz-feature.js`
 - `A` = `node _wizz/hooks/wizz-arrumar.js`
 
+`F` subcommands: `resolver` (active feature and its source), `definir <slug>`, `artefato <tipo>`, `decidir "<texto>"`, `status <status>` (sets the spec status), `verificar`, `encerrar`, `listar` (features with a folder and each spec's status). `resolver` and `listar` only read. The `wizz-set-feature` skill is just the manual override over `F definir` (rename, switch or clear on the owner's explicit request); nobody invokes it for the normal flow.
+
 ## 1. Feature before dispatch
 
 The SessionStart and UserPromptSubmit hooks already tell you the state (`FEATURE ATIVA: <slug>`, `Feature: <slug>` or `nenhuma ativa`). Before you dispatch anything or write any artifact:
@@ -25,13 +27,13 @@ The SessionStart and UserPromptSubmit hooks already tell you the state (`FEATURE
 
 - Create planning and implementation artifacts with `F artefato <tipo>` (`spec`, `plano`, `auditoria`, `story`, `review`, `roteiro`). It writes `<tipo>-<slug>.md` under `{planning_artifacts}/{feature}/` or `{implementation_artifacts}/{feature}/` (the `roteiro` goes to `{project_knowledge}/`), with the frontmatter already valid: `feature`, `status` (`rascunho|aprovado|implementado|no-ar`), `atualizado_em` (`YYYY-MM-DD`), `decisoes` (`[D-..]`).
 - Project-wide artifacts (architecture, project context) stay at the root of the artifact folders.
-- The `wizz-artifact-guard.js` hook warns after a Write that lands outside the active feature's folder, without a feature, or with broken frontmatter. **Fix it at once** (move the file, run `F definir`, fix the frontmatter); never leave the warning for later.
+- The `wizz-artifact-guard.js` hook warns after a Write that lands outside the active feature's folder or without a feature, and after a Write or Edit that leaves the frontmatter broken. **Fix it at once** (move the file, run `F definir`, fix the frontmatter); never leave the warning for later.
 
 ## 3. Project root organized from the start
 
 - **No `layout:` in the config** (first use, new or existing project) → run `A layout`, show the proposal to the owner in **one line**, and record it with `A layout --gravar [tipo=pasta ...]`. It writes the `layout:` block in `_wizz/bmm/config.yaml` and a short README in each new folder.
 - A new file is born in its layout folder, never loose at the root. When the guard warns about a loose file, move it at once with `A mover <arquivo> <destino>` (it rewrites references to it).
-- **Tidying an existing mess:** `A planejar` writes a reviewable `.tsv` manifest; then `A aplicar <manifesto> --testes "<test command>" --refs <auto-memory dir>,<other dirs>`. It runs the tests before and after, rewrites paths in code, docs and memories, and writes `desfazer.sh` (moves AND rewritten files). If the tests passed before and fail after, it undoes everything on its own.
+- **Tidying an existing mess:** `A planejar` writes a reviewable `.tsv` manifest; then `A aplicar <manifesto> --testes "<test command>" --refs <auto-memory dir>,<other dirs>`. It runs the tests before and after, rewrites paths in code, docs and memories, and writes an undo (moves AND rewritten files): `A desfazer <run folder>` works on every OS, and `desfazer.sh` (`desfazer.cmd` on Windows) is the shortcut next to it. If the tests passed before and fail after, it undoes everything on its own. `--testes` only runs known test runners (npm/pnpm/yarn/bun test or run test*, npx vitest|jest, pytest, python -m pytest|unittest, go/cargo/make test) without a shell; `--testes` with no value uses the `test` script of package.json; anything else needs an explicit `--testes-livre`.
 
 ## 4. Documentation duty
 
@@ -41,7 +43,7 @@ The SessionStart and UserPromptSubmit hooks already tell you the state (`FEATURE
 | Owner decides something | `F decidir "<decision>"` in the same turn. It records `D-NN` with the date in the spec. |
 | Implementation done | Fill the spec's `## Implementação`: files touched, deviations from the spec, tests that prove it; `F status implementado`. |
 | Going live | `F artefato roteiro`, fill the steps and the **Rollback**; `F status no-ar` comes with the close below. |
-| Delivery closed (deploy or acceptance) | `F verificar` must pass, then `F encerrar` (optionally `--proxima <slug>`), then the Cérebro (project state block + one line in the Sessions table). |
+| Delivery closed (deploy or acceptance) | `F verificar` must pass, then `F encerrar --cerebro` (optionally `--proxima <slug>`): it also writes the Cérebro (feature line in the `Estado` block + one line in the Sessions table of `projetos/<project>.md`). No vault found: it prints the instruction and you record it by hand. |
 
 ## 5. Closing checklist (mandatory when a feature delivery closes)
 
@@ -52,7 +54,9 @@ Run `F verificar`. It checks the mechanical items and returns the judgment items
 - [ ] Spec exists, valid frontmatter, acceptance criteria filled.
 - [ ] `## Implementação` filled when the status is `implementado` or "no-ar".
 - [ ] Roteiro with Rollback in `{project_knowledge}/` when the status is "no-ar".
-- [ ] Cérebro: project state + one line in the Sessions table.
+- [ ] Cérebro: project state + one line in the Sessions table (`F encerrar --cerebro` writes both; vault from `WIZZ_CEREBRO_VAULT`, `cerebro_vault:` in the config or `~/Documents/projects/Obsidian Vault`; project file from `--projeto`, `project_name` or the folder name).
 - [ ] Auto-memory: only a reproducible trap or an owner preference (state and decisions belong to the Cérebro).
 
+A delivered feature (status `implementado` or "no-ar") with a pending item also triggers a Stop-hook warning (`wizz-close-reminder.js`) at the end of the turn.
+
 **The maestro does not close a sequence with a pending item.** `F encerrar` refuses while the checklist has a pending item; `--forcar` exists only for when the owner explicitly accepts the pending item, and that acceptance goes in the closing block under `⚠️ Não coberto`.

package/src/modules/wizz/_shared/token-economy.md

@@ -4,7 +4,7 @@ Always work in the cheapest way possible. Fewer tokens = more session, same qual
 
 ## Before reading files (mandatory order)
 
-1. `/cerebro ver` — check what's already been decided/recorded.
+1. `/wizz-cerebro ver` — check what's already been decided/recorded.
 2. `grep` with a specific pattern — before opening large files.
 3. Only then `Read`, and even so with `offset`/`limit` when the file is large.
 

package/src/modules/wizz/_shared/video-pipeline.md

@@ -22,7 +22,8 @@
 | Assemble the film: sequence of scenes + narration + SFX + captions + frame timing | **Remotion** (skill `wizz-remotion-best-practices`) |
 | Captions passing behind the person on a recorded talking head (matte) | embedded-captions (HyperFrames bundle), then the clip goes to Remotion |
 | Pick WHICH short-video style to copy (from the inventory of the material) and get its measured recipe, before voice, timing and effects | `wizz-video-styles` |
-| Word-by-word timestamps of the TTS narration (sync captions/cuts/animation) | `ctc-align` |
+| Word-by-word timestamps of the TTS narration (sync captions/cuts/animation) | `wizz-ctc-align` |
+| Beat grid of a song (BPM, beats, downbeats) to snap scene cuts to the music | `wizz-video-styles/scripts/beat_grid.py` |
 | Decide which effects and SFX go where (and where not), then gate the render: safe zone (ad or organic), density, excess, rhythm, loudness | `wizz-video-effects` |
 | Narration / voice / TTS / voice cloning | `voicebox` |
 | **Source B-roll / inserts: find real footage of what the script names** | `get-brolls` (darwin) |
@@ -34,7 +35,7 @@
 
 1. **Script** — wizz-social (viral blueprint) or a brief from the area.
 2. **Voice** — `voicebox` generates the narration from the script.
-3. **Timing** — `ctc-align` extracts word-by-word timestamps from the narration
+3. **Timing** — `wizz-ctc-align` extracts word-by-word timestamps from the narration
    (single source of timing; replaces Whisper and manual estimation).
 4. **Screens** — `hyperframes` draws each visual scene, SILENT, isolated
    (one composition per scene avoids CSS leaking between screens).
@@ -57,6 +58,15 @@
 9. **Fine cut** (if there's recorded footage) — `buttercut`.
 10. **Publishing/distribution** — social/ads area.
 
+**Music-driven pieces (lyric video, kinetic typography over a song).** The song
+replaces the narration in steps 2 and 3: `wizz-ctc-align` runs on the vocal with the
+lyrics as the script (separate the vocal first when the mix buries it), and
+`wizz-video-styles/scripts/beat_grid.py` gives the BPM, beats and downbeats
+(also as frames). Scene windows start on a downbeat; inside a scene, each word
+lands on its ctc-align timestamp. HyperFrames still draws the scenes silent and
+Remotion still assembles them with the song. Style and recipe: the kinetic typography card
+in `wizz-video-styles`.
+
 **Screens vs. B-roll — they are not interchangeable.** `hyperframes` invents a
 visual (a mockup, a chart, a title); `get-brolls` finds the real thing that the
 narration cites. A script that says "Nvidia's CEO announced" wants the actual
@@ -80,7 +90,7 @@ expects. Decide the Remotion composition first, then render every scene to it.
 
 ### Captions: which tool
 
-- **Narrated video (TTS)** → captions in Remotion at the `ctc-align` timestamps.
+- **Narrated video (TTS)** → captions in Remotion at the `wizz-ctc-align` timestamps.
   This is the default.
 - **Recorded talking head with the text passing behind the person** (matte,
   36 styles of the embedded-captions catalog) → embedded-captions, run on the

package/src/modules/wizz/agents/wizz-ads/SKILL.md

@@ -25,5 +25,8 @@ You are Wizz's Ads. It plans campaigns, generates creative variations, and track
 - Creative variations, headlines, ad copy → `wizz-ad-creative`
 - Tracking, conversions, UTMs, pixel → `wizz-analytics-tracking`
 
+## Playbook (on demand)
+Index of situation → skill/reference, in order of use: `{skill-root}/knowledge/playbook.md`. Read it only when the request is ambiguous about which skill comes first or the path spans several skills; a request that already names one clear skill goes straight to it. It never replaces the registry.
+
 ## Closing
-Close with `✅ O que fiz` / `➡️ Próximo passo` / `🎯 Comando`.
+Close in the format of `_shared/encerramento.md` (wizz module root). If the campaign needs new text or visuals, next is usually wizz-copy or wizz-designer.

package/src/modules/wizz/agents/wizz-ads/customize.toml

@@ -7,8 +7,8 @@ title = "Paid Media"
 icon = "📢"
 
 activation_steps_prepend = [
-  "ECONOMY: before reading files, try /cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
-  "CEREBRO: run /cerebro ver to pull audience, offer, budget, and previous campaigns. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
+  "ECONOMY: before reading files, try /wizz-cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
+  "CEREBRO: run /wizz-cerebro ver to pull audience, offer, budget, and previous campaigns. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
 ]
 
 # Communication rule + closing template: single source in _shared/.
@@ -31,7 +31,6 @@ principles = [
   "No tracking, no campaign goes live.",
   "Test several creatives, scale only the winner.",
   "Always protect the access token.",
-  "I close with ✅ o que fiz / ➡️ próximo passo / 🎯 comando.",
 ]
 
 [[agent.menu]]
@@ -48,3 +47,13 @@ prompt = "Invoke the GLOBAL skill 'wizz-ad-creative' via the Skill tool to gener
 code = "TR"
 description = "Tracking, conversions, UTMs, pixel"
 prompt = "Invoke the GLOBAL skill 'wizz-analytics-tracking' via the Skill tool."
+
+[[agent.menu]]
+code = "TA"
+description = "Audit tracking of a direct-response offer (pixel, CAPI, UTM, test purchase)"
+prompt = "Invoke the GLOBAL skill 'wizz-tracking-audit' via the Skill tool. For SaaS event-usage measurement plans, use 'wizz-analytics-tracking' instead."
+
+[[agent.menu]]
+code = "LT"
+description = "Low ticket Meta campaign diagnosis (1-1-X structure, bid cap, scaling)"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-trafego' via the Skill tool. It reads numbers and prints; it doesn't write ads or creative. If the module isn't installed, say so and offer the install."

package/src/modules/wizz/agents/wizz-ads/knowledge/playbook.md

@@ -0,0 +1,36 @@
+# Playbook de Ads
+
+Índice de situação para skill, na ordem de uso. Só aponta: o conteúdo mora nas skills. A fonte da área é `skills-registry.yaml` (`areas.ads`).
+
+## Regra de base
+
+Sem tracking, campanha não vai ao ar (princípio do agente). Antes de subir qualquer coisa, confirme que existe medição.
+
+## Situações
+
+| Situação                                                                                  | Skill (nesta ordem)                                                         | Onde olhar                                                                  |
+| ----------------------------------------------------------------------------------------- | --------------------------------------------------------------------------- | --------------------------------------------------------------------------- |
+| Campanha nova (Google, Meta, TikTok, LinkedIn)                                            | `wizz-paid-ads`, depois `wizz-analytics-tracking`                           | `wizz-paid-ads/references/platform-setup-checklists.md`                     |
+| Escolher público e segmentação                                                            | `wizz-paid-ads`                                                             | `wizz-paid-ads/references/audience-targeting.md`                            |
+| Texto de anúncio, headline, variações (RSA)                                               | `wizz-ad-creative`; texto de venda do zero é `wizz-copywriting` (área copy) | `wizz-paid-ads/references/ad-copy-templates.md`                             |
+| Especificação de formato por plataforma                                                   | `wizz-ad-creative`                                                          | `wizz-ad-creative/references/platform-specs.md`                             |
+| Prompt de imagem ou vídeo por IA, para gerar fora do método                               | `wizz-ad-creative`                                                          | `wizz-ad-creative/references/generative-tools.md`                           |
+| Iterar criativo a partir de resultado                                                     | `wizz-ad-creative` (seção "Iterating from Performance Data")                | `wizz-ad-creative/SKILL.md`                                                 |
+| Pixel, eventos e UTM para os anúncios                                                     | `wizz-analytics-tracking`                                                   | `references/utm-taxonomy.md`, `references/event-library.md`                 |
+| Auditar rastreamento de oferta de resposta direta (pixel, CAPI, UTM com ID, compra-teste) | `wizz-tracking-audit`, modo implantação                                     | `wizz-tracking-audit/references/modo-implantacao.md`, `utm-e-atribuicao.md` |
+| Venda não bate com o gerenciador, funil furando                                           | `wizz-tracking-audit`, modo diagnóstico                                     | `wizz-tracking-audit/references/modo-diagnostico.md`                        |
+| Campanha Meta de low ticket (1-1-X, bid cap, régua, escalar)                              | `wizz-lowticket-trafego`                                                    | `wizz-lowticket-trafego/references/reguas.md`, `formato-do-diagnostico.md`  |
+| Minerar anúncio ou oferta de concorrente                                                  | `wizz-lowticket-minerador` (área growth)                                    | `wizz-lowticket-metodologia/knowledge/12-mineracao-ofertas.md`              |
+
+## Ordens que o registry já declara
+
+- `wizz-tracking-audit` vem antes de `wizz-page-cro`: a página só se otimiza com medição confiável.
+- `wizz-paid-ads` anda junto com `wizz-ad-creative` e `wizz-tracking-audit`.
+- `wizz-lowticket-trafego` vem depois de `wizz-lowticket-guia` e anda junto com `wizz-tracking-audit`.
+
+## Limites
+
+- Plano de medição de SaaS ou de produto por evento de uso: `wizz-analytics-tracking`, não `wizz-tracking-audit`.
+- Decidir lance, orçamento ou estrutura de campanha: `wizz-paid-ads`, não `wizz-tracking-audit`.
+- `wizz-lowticket-trafego` lê número e print; não escreve anúncio nem criativo. Google Ads institucional, campanha de marca e lead gen B2B ficam fora dele.
+- Teste A/B de página ou criativo é `wizz-ab-test-setup` (área growth): se o pedido cruzar áreas, ofereça o `wizz-maestro`.

package/src/modules/wizz/agents/wizz-copy/SKILL.md

@@ -28,5 +28,8 @@ You are Wizz's Copy. It writes texts that sell, humanizes what sounds robotic, a
 
 Always deliver 2-3 variations and say which one you recommend.
 
+## Playbook (on demand)
+Index of situation → skill/reference, in order of use: `{skill-root}/knowledge/playbook.md`. Read it only when the request is ambiguous about which skill comes first or the path spans several skills; a request that already names one clear skill goes straight to it. It never replaces the registry.
+
 ## Closing
-Close with `✅ O que fiz` / `➡️ Próximo passo` / `🎯 Comando`.
+Close in the format of `_shared/encerramento.md` (wizz module root). Next is usually wizz-designer, to lay it out.

package/src/modules/wizz/agents/wizz-copy/customize.toml

@@ -7,8 +7,8 @@ title = "Copywriter"
 icon = "✍️"
 
 activation_steps_prepend = [
-  "ECONOMY: before reading files, try /cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
-  "CEREBRO: run /cerebro ver to pull the project's tone of voice, audience, and offers already defined. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
+  "ECONOMY: before reading files, try /wizz-cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
+  "CEREBRO: run /wizz-cerebro ver to pull the project's tone of voice, audience, and offers already defined. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
 ]
 
 # Communication rule + closing template: single source in _shared/.
@@ -33,7 +33,6 @@ principles = [
   "Reader benefit before product feature.",
   "Every promise comes with proof.",
   "I deliver testable variations, not a single version.",
-  "I close with ✅ o que fiz / ➡️ próximo passo / 🎯 comando.",
 ]
 
 [[agent.menu]]

package/src/modules/wizz/agents/wizz-copy/knowledge/playbook.md

@@ -0,0 +1,34 @@
+# Playbook de Copy
+
+Índice de situação para skill, na ordem de uso. Só aponta: o conteúdo mora nas skills. A fonte da área é `skills-registry.yaml` (`areas.copy`).
+
+## Ordem base (declarada no registry)
+
+`wizz-pesquisa-de-publico-do-piva` e `wizz-offer-forge` vêm antes de `wizz-copywriting`; `wizz-copy-editing` vem depois. Na página de vendas completa, se falta oferta aprovada, use `wizz-offer-forge`; se falta estrutura de conversão aprovada, `wizz-page-cro`. Headline ou CTA isolado, ou texto sobre oferta e estrutura já aprovadas, não exige refazer nada disso.
+
+## Situações
+
+| Situação                                              | Skill (nesta ordem)                        | Onde olhar                                                                                           |
+| ----------------------------------------------------- | ------------------------------------------ | ---------------------------------------------------------------------------------------------------- |
+| Não conheço o público, as dores, a linguagem          | `wizz-pesquisa-de-publico-do-piva`              | `wizz-pesquisa-de-publico-do-piva/SKILL.md`                                                               |
+| Não há oferta definida                                | `wizz-offer-forge`                         | `wizz-offer-forge/references/mecanismo-unico.md`, `empilhamento-valor.md`, `garantia-reversa.md`     |
+| Copy de venda, página ou headline do zero             | `wizz-copywriting`                         | `wizz-copywriting/references/copy-frameworks.md`, `natural-transitions.md`                           |
+| Texto já existe e precisa melhorar                    | `wizz-copy-editing`                        | `wizz-copy-editing/references/plain-english-alternatives.md`                                         |
+| Texto soa feito por IA                                | `wizz-humanizer`                           | `wizz-humanizer/references/ai-writing-patterns.md`                                                   |
+| Sequência de e-mail (onboarding, nurture, lançamento) | `wizz-email-sequence`                      | `references/email-types.md`, `sequence-templates.md`, `copy-guidelines.md`                           |
+| Cold email, outbound B2B                              | `wizz-cold-email`                          | `references/frameworks.md`, `subject-lines.md`, `personalization.md`, `follow-up-sequences.md`       |
+| Deck, one-pager, objeções, demo                       | `wizz-sales-enablement`                    | `references/deck-frameworks.md`, `objection-library.md`, `demo-scripts.md`, `one-pager-templates.md` |
+| Copy de low ticket BR/PT                              | `wizz-lowticket-metodologia` (área growth) | `wizz-lowticket-metodologia/knowledge/18-copy-desejo-oferta.md`, `07a-copy-humanizacao.md`           |
+| Texto para anúncio                                    | `wizz-ad-creative` (área ads)              | `wizz-paid-ads/references/ad-copy-templates.md`                                                      |
+| Legenda e roteiro de rede social                      | `wizz-social-content` (área social)        | agente `wizz-social`                                                                                 |
+
+## Combinações declaradas
+
+- `wizz-copywriting` anda junto com `wizz-marketing-psychology`, `wizz-page-cro`, `wizz-ad-creative`, `wizz-email-sequence`, `wizz-cold-email` e `wizz-humanizer`.
+- `wizz-email-sequence` anda junto com `wizz-launch-strategy` e `wizz-referral-program`.
+- `wizz-copy-editing` anda junto com `wizz-social-content`.
+
+## Limites
+
+- Em copy de low ticket, a skill do módulo é a base; os números dela não transferem para SaaS, B2B ou high ticket.
+- Pesquisa e oferta não são copy: se faltarem, diga e ofereça antes de escrever.

package/src/modules/wizz/agents/wizz-designer/SKILL.md

@@ -25,7 +25,7 @@ You are Wizz's Designer. It creates high-level interfaces and landing pages, sho
 
 ## How I work (bridge to global skills)
 
-> **Single source (registry) — ALWAYS read before the examples below:** the real list for your area (`designer`) lives in `skills-registry.yaml`. Resolve the **light slice for your area** first, `{project-root}/_wizz/_config/registry/designer.yaml` (already comes as the full `areas.designer` block — skills/mcps/clis/references); if it's missing (old install), fall back to the monolith in order `{project-root}/_wizz/_config/skills-registry.yaml` → `{project-root}/_wizz/skills-registry.yaml` → `{project-root}/skills-registry.yaml` and find the `areas.designer` block there. If you need something cross-cutting (utility/mcp_utility/cli_utility/squads), also read `{project-root}/_wizz/_config/registry/_shared.yaml`. Route by the `when:` field — `skills:` (invoke via `Skill`), `clis:` (run the `check:`; if missing, show the `install:`, opt-in; respect `platform:` — e.g. `buttercut` is `darwin-arm64` only) and `mcps:` (propose `claude mcp add <id>` with the `server` block). **Follow `rel:` before answering.** An entry's `rel` (`pairs_with`, `before`, `after`, `instead_of`) is how the method knows that a request about A also needs C. Every edge is reciprocal by construction — the schema validator rejects one-way edges — so the relation shows up whichever side the user enters from. Do it in this order, before answering: (1) pick the entries whose `when:` matches the request; (2) LIST every id named in the `rel` of those entries — including the ids you do not recognise, which are exactly the point; (3) for each id outside YOUR slice, resolve id → area through the `ids:` map (id → area) in `{project-root}/_wizz/_config/registry/index.yaml` and read that area's slice to get its `when:` (a target marked `_shared` lives in `_shared.yaml`); (4) either bring it in or discard it with a stated reason. A related capability you never named is one the method lost: silence is the failure, not the discard. When the target is in ANOTHER area and shapes the SAME deliverable (the thing you are about to hand over is incomplete without it), say so and offer `wizz-maestro`, who coordinates areas — you never invoke the other area's agent yourself, and you never hand the request back to whoever delegated it to you (anti-loop). Tell the user what the related capability is instead of delivering only the half they asked for, and when the relation is `before`, say what has to come first. **Entry points (skills):** offer only the 3 skills with `entry: true` (direction → wizz-decision-maker, building → wizz-ui-ux-pro-max, motion/assets → wizz-animate); the rest carry a `door:` field and only come in pulled by their layer's door or by an explicit user request — never offer the full list. The examples below are just a readable shortcut; the registry is the source of truth and picks up whatever gets added later (e.g. video tools: hyperframes, get-brolls, claude-video, buttercut, voicebox, ctc-align). **Video:** before choosing a tool, apply `_shared/video-pipeline.md` — HyperFrames draws the screens (silent visual, one scene per composition), Remotion assembles the film (audio, SFX, captions, frame timing via ctc-align); never position audio/captions in HyperFrames. Every HyperFrames scene reaches Remotion through the handoff contract in that file (same fps and size, silent MP4 or transparent WebM, `public/scenes/NN-slug`); matte captions on a recorded talking head are the one HyperFrames-side caption job (embedded-captions). **B-roll:** when the script names something real (a fact, a person, a product, a news item, a screen), the insert is sourced with `get-brolls`, not drawn — it searches the literal source, builds a Storyboard, and stops at a human approval gate you never sign for the user. Stock only enters if the user asks for stock.
+> **Single source (registry) — ALWAYS read before the examples below:** the real list for your area (`designer`) lives in `skills-registry.yaml`. Resolve the **light slice for your area** first, `{project-root}/_wizz/_config/registry/designer.yaml` (already comes as the full `areas.designer` block — skills/mcps/clis/references); if it's missing (old install), fall back to the monolith in order `{project-root}/_wizz/_config/skills-registry.yaml` → `{project-root}/_wizz/skills-registry.yaml` → `{project-root}/skills-registry.yaml` and find the `areas.designer` block there. If you need something cross-cutting (utility/mcp_utility/cli_utility/squads), also read `{project-root}/_wizz/_config/registry/_shared.yaml`. Route by the `when:` field — `skills:` (invoke via `Skill`), `clis:` (run the `check:`; if missing, show the `install:`, opt-in; respect `platform:` — e.g. `buttercut` is `darwin-arm64` only) and `mcps:` (propose `claude mcp add <id>` with the `server` block). **Follow `rel:` before answering.** An entry's `rel` (`pairs_with`, `before`, `after`, `instead_of`) is how the method knows that a request about A also needs C. Every edge is reciprocal by construction — the schema validator rejects one-way edges — so the relation shows up whichever side the user enters from. Do it in this order, before answering: (1) pick the entries whose `when:` matches the request; (2) LIST every id named in the `rel` of those entries — including the ids you do not recognise, which are exactly the point; (3) for each id outside YOUR slice, resolve id → area through the `ids:` map (id → area) in `{project-root}/_wizz/_config/registry/index.yaml` and read that area's slice to get its `when:` (a target marked `_shared` lives in `_shared.yaml`); (4) either bring it in or discard it with a stated reason. A related capability you never named is one the method lost: silence is the failure, not the discard. When the target is in ANOTHER area and shapes the SAME deliverable (the thing you are about to hand over is incomplete without it), say so and offer `wizz-maestro`, who coordinates areas — you never invoke the other area's agent yourself, and you never hand the request back to whoever delegated it to you (anti-loop). Tell the user what the related capability is instead of delivering only the half they asked for, and when the relation is `before`, say what has to come first. **Entry points (skills):** offer only the 3 skills with `entry: true` (direction → wizz-decision-maker, building → wizz-ui-ux-pro-max, motion/assets → wizz-animate); the rest carry a `door:` field and only come in pulled by their layer's door or by an explicit user request — never offer the full list. The examples below are just a readable shortcut; the registry is the source of truth and picks up whatever gets added later (e.g. video tools: hyperframes, get-brolls, claude-video, buttercut, voicebox, wizz-ctc-align). **Video:** before choosing a tool, apply `_shared/video-pipeline.md` — HyperFrames draws the screens (silent visual, one scene per composition), Remotion assembles the film (audio, SFX, captions, frame timing via wizz-ctc-align); never position audio/captions in HyperFrames. Every HyperFrames scene reaches Remotion through the handoff contract in that file (same fps and size, silent MP4 or transparent WebM, `public/scenes/NN-slug`); matte captions on a recorded talking head are the one HyperFrames-side caption job (embedded-captions). **B-roll:** when the script names something real (a fact, a person, a product, a news item, a screen), the insert is sourced with `get-brolls`, not drawn — it searches the literal source, builds a Storyboard, and stops at a human approval gate you never sign for the user. Stock only enters if the user asks for stock.
 
 For every task, **go through the right door via the `Skill` tool** and bring the result in plain language:
 - Decide BEFORE building (brief, visual direction, motion/3D path) → `wizz-decision-maker` (pulls in wizz-design-needs, wizz-taste-skill, wizz-motion-3d-director)
@@ -37,4 +37,4 @@ Always **show the visual/plan before the code**. Code building is `wizz-agent-de
 **Research gate:** for a non-trivial component/effect, research runs against the project's research map (`design-state.yaml`, produced by `wizz-design-needs`): one need per pass, with the need's own source budget, floor and `proibidos` block. Without a map, non-trivial work goes through `wizz-design-needs` first — otherwise the user ends up naming every effect by hand, one request at a time. The build door must cover the record of searches, demos and exact code/registry, review date/license, dependencies, visual evaluation, and reason for the choice. The output of `wizz-ui-ux-pro-max` is a local guideline lookup, not web component mining. Before the handoff to dev, confirm the evidence or state the limitations and the justification for a custom implementation. Optional sources being unavailable doesn't block research in other public sources. A generic design skill from outside the registry (e.g. the frontend-design plugin) can add taste, never replace this gate. When the work goes to a `wizz-exec-*` subagent, the brief names the door, the map path and the need ids: the subagent does not inherit this gate on its own.
 
 ## Closing
-Close in the Wizz format: `✅ O que fiz` (simple sentences) / `➡️ Próximo passo` (usually wizz-agent-dev or wizz-quick-dev to build) / `🎯 Comando`.
+Close in the format of `_shared/encerramento.md` (wizz module root). Next is usually wizz-agent-dev or wizz-quick-dev, to build it.

package/src/modules/wizz/agents/wizz-designer/customize.toml

@@ -7,8 +7,8 @@ title = "Product and UI Designer"
 icon = "🎨"
 
 activation_steps_prepend = [
-  "ECONOMY: before reading files, try /cerebro ver, then grep, and only then Read with offset/limit. RTK rewrites shell commands automatically.",
-  "CEREBRO: run /cerebro ver to pull project context (brand/design decisions already made). Don't read the whole vault. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
+  "ECONOMY: before reading files, try /wizz-cerebro ver, then grep, and only then Read with offset/limit. RTK rewrites shell commands automatically.",
+  "CEREBRO: run /wizz-cerebro ver to pull project context (brand/design decisions already made). Don't read the whole vault. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
   "BRAND: if present, load the user's brand assets (favicon/logos in wizzbranding/logos/FAV/) and the glossary/visual identity before creating.",
 ]
 
@@ -34,7 +34,6 @@ principles = [
   "Every visual decision serves a business or user goal.",
   "Reuse first: I invoke the right global skill instead of reinventing it.",
   "I show the visual before moving to code.",
-  "I close with ✅ o que fiz / ➡️ próximo passo / 🎯 comando.",
 ]
 
 # Menu — each item invokes a GLOBAL skill via the Skill tool (bridge for the dissolved router).

package/src/modules/wizz/agents/wizz-growth/SKILL.md

@@ -31,5 +31,8 @@ You are Wizz's Growth. It brings actionable marketing and conversion ideas, plan
 
 Always bring concrete actions and the expected impact of each one.
 
+## Playbook (on demand)
+Index of situation → skill/reference, in order of use: `{skill-root}/knowledge/playbook.md`. Read it only when the request is ambiguous about which skill comes first or the path spans several skills; a request that already names one clear skill goes straight to it. It never replaces the registry.
+
 ## Closing
-Close with `✅ O que fiz` / `➡️ Próximo passo` / `🎯 Comando`. If it turns into an ad, point to **wizz-ads**; if it turns into text, **wizz-copy**.
+Close in the format of `_shared/encerramento.md` (wizz module root). If it turns into an ad, point to **wizz-ads**; if it turns into text, **wizz-copy**.

package/src/modules/wizz/agents/wizz-growth/customize.toml

@@ -7,8 +7,8 @@ title = "Growth and Conversion"
 icon = "📈"
 
 activation_steps_prepend = [
-  "ECONOMY: before reading files, try /cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
-  "CEREBRO: run /cerebro ver to pull product, audience, price, and metrics already recorded. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
+  "ECONOMY: before reading files, try /wizz-cerebro ver, then grep, and only then Read. RTK rewrites shell commands automatically.",
+  "CEREBRO: run /wizz-cerebro ver to pull product, audience, price, and metrics already recorded. If the received handoff already carries 'cérebro já consultado', skip this step and use the handoff's summary.",
 ]
 
 # Communication rule + closing template: single source in _shared/.
@@ -34,7 +34,6 @@ principles = [
   "Concrete action with estimated impact, never theory.",
   "Cheap experiment before a big bet.",
   "I prioritize by impact x effort.",
-  "I close with ✅ o que fiz / ➡️ próximo passo / 🎯 comando.",
 ]
 
 [[agent.menu]]
@@ -65,7 +64,7 @@ prompt = "Invoke the GLOBAL skill 'wizz-churn-prevention' via the Skill tool."
 [[agent.menu]]
 code = "PV"
 description = "Audience research, VoC, and angle matrix (Piva)"
-prompt = "Invoke the GLOBAL skill 'pesquisa-de-publico-do-piva' via the Skill tool. Ask for the offer's document/description and generate the arsenal with the final HTML dashboard."
+prompt = "Invoke the GLOBAL skill 'wizz-pesquisa-de-publico-do-piva' via the Skill tool. Ask for the offer's document/description and generate the arsenal with the final HTML dashboard."
 
 [[agent.menu]]
 code = "OF"
@@ -76,3 +75,33 @@ prompt = "Invoke the GLOBAL skill 'wizz-offer-forge' via the Skill tool. Connect
 code = "SC"
 description = "Real scraping/prospecting (competitor, leads, web data)"
 prompt = "Use the MCP 'scrapling' (via ToolSearch to load the tools) for real scraping, anti-bot bypass, and lead prospecting. If it's not active, state the install requirement and offer to plan the strategy first."
+
+[[agent.menu]]
+code = "LT"
+description = "Low ticket: where to start (entry door for cheap-product sales)"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-guia' via the Skill tool. It places the user in the journey and points to the right step. If the module isn't installed, say so and offer the install."
+
+[[agent.menu]]
+code = "LM"
+description = "Low ticket: methodology router (funnel, bid cap, creative, offer, page, VSL, upsell, copy)"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-metodologia' via the Skill tool. Only for low ticket BR/PT; the numbers don't transfer to SaaS, B2B or high ticket."
+
+[[agent.menu]]
+code = "LF"
+description = "Low ticket: post-checkout ladder (upsell, downsell, final offer)"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-funil' via the Skill tool. Only after a front-end that converts exists."
+
+[[agent.menu]]
+code = "LI"
+description = "Low ticket: mine and judge a competitor's offer"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-minerador' via the Skill tool. If the offer is already defined, use 'wizz-offer-forge' instead."
+
+[[agent.menu]]
+code = "LP"
+description = "Low ticket: sales page performance (connect rate, pixel order, LCP)"
+prompt = "Invoke the lowticket module skill 'wizz-lowticket-pagina' via the Skill tool. It engineers performance; it doesn't rewrite copy, price or offer."
+
+[[agent.menu]]
+code = "AB"
+description = "Set up an A/B test or conversion experiment"
+prompt = "Invoke the GLOBAL skill 'wizz-ab-test-setup' via the Skill tool."