@etus/seven-skill 0.1.0-beta.3 → 0.1.0-beta.4

package/.kiro/skills/seven/reference/audit.md

@@ -0,0 +1,122 @@
+# Audit
+
+Run systematic technical quality checks on a Seven surface or component and generate a report. The audit documents issues; it does not fix them. Hand findings to `polish` or to direct implementation.
+
+## Scope
+
+Code-level audit, not design critique. Check what is measurable and verifiable against `@etus/ui`, `@etus/tokens`, and Seven's documented rules.
+
+## Five-dimension diagnostic scan
+
+Score each dimension 0-4 using the criteria below. The detector ships rule IDs you can cite directly.
+
+### 1. Accessibility (a11y)
+
+Check for:
+
+- Text contrast ratios below 4.5:1 (or 3:1 for large text). Verify against the OKLCH math from `@etus/tokens`, not eyeballed.
+- Missing ARIA roles, labels, or states on interactive elements built from non-semantic primitives.
+- Keyboard traps, missing focus indicators, illogical tab order.
+- Improper heading hierarchy, missing landmarks, `<div onClick>` instead of `<button>`.
+- Image and icon-only buttons without accessible names.
+- Forms with inputs missing labels, error messages missing `aria-describedby`, required state not announced.
+- `prefers-reduced-motion` not honored.
+
+Scoring: 0 = fails WCAG A, 1 = major gaps, 2 = partial, 3 = WCAG AA mostly met, 4 = AA fully met and approaching AAA.
+
+### 2. Performance
+
+Check for:
+
+- Layout thrashing in loops.
+- Animation on CSS layout properties (`width`, `height`, `top`, `left`). Transform and opacity only.
+- Unbounded blur, filter, shadow effects that visibly drop frames.
+- Images without `loading="lazy"`, unoptimized assets, missing dimensions.
+- Unnecessary re-renders, missing `useMemo` or `React.memo` where measurement shows benefit (do not memoize speculatively).
+- Heavy imports pulled in eagerly when a dynamic import would do.
+
+Scoring: 0 = severe issues, 1 = major problems, 2 = partial optimization, 3 = mostly optimized, 4 = lean and fast.
+
+### 3. Theming
+
+Check for:
+
+- Hardcoded color literals in runtime code. Hex inside `packages/tokens/sources/` is fine; anywhere else is drift. Detector rule: `hardcoded-color-not-token`.
+- Pure `#000` and `#fff`. Detector rule: `pure-black-or-white`.
+- Tailwind v4 arbitrary values referencing color tokens without the `color:` hint. Detector rule: `missing-color-type-hint`.
+- Font-size or line-height arbitrary values without `length:`. Detector rule: `missing-length-type-hint`.
+- `leading-none` combined with a typed length-variable text utility. Detector rule: `leading-none-with-length-var`.
+- Wrong token used: `var(--background)` where `var(--muted)` would have read better; mixing token families.
+- Dark mode visually broken: tokens that look fine in light fail at AAA contrast in dark, or vice versa.
+
+Scoring: 0 = no token usage, 1 = minimal tokens, 2 = partial token coverage with gaps, 3 = good token usage with minor literals, 4 = full token system with verified dark mode.
+
+### 4. Responsive design
+
+Check for:
+
+- Fixed widths that break on narrow viewports.
+- Touch targets below 44 x 44.
+- Horizontal scroll on mobile.
+- Text scaling: layout breaks when the user bumps base font size.
+- Missing or wrong-direction breakpoints (mobile-first vs desktop-first inconsistency in the same component).
+
+Scoring: 0 = desktop-only, 1 = major issues, 2 = partial, 3 = good with minor issues, 4 = fluid across viewports.
+
+### 5. Anti-patterns
+
+Run the deterministic detector against the surface or component:
+
+```bash
+seven check --file <path>
+```
+
+Then check the report against the Seven rule set:
+
+| Detector rule ID               | Catches                                                                 |
+| ------------------------------ | ----------------------------------------------------------------------- |
+| `non-inter-font`               | Font family other than Inter or JetBrains Mono in consumer code         |
+| `hardcoded-color-not-token`    | Hex, rgb, hsl, OKLCH literal where a token exists                       |
+| `missing-color-type-hint`      | Tailwind v4 arbitrary color value missing the `color:` hint             |
+| `missing-length-type-hint`     | Tailwind v4 arbitrary length value missing the `length:` hint           |
+| `leading-none-with-length-var` | `leading-none` combined with a typed length-variable text utility       |
+| `fuchsia-in-ui-chrome`         | Fuchsia color on UI chrome (allowed on `data-role="chart"` and `<svg>`) |
+| `pt-pt-vocab`                  | European Portuguese vocabulary in pt-BR surfaces                        |
+| `pure-black-or-white`          | `#000`, `#fff`, `rgb(0,0,0)`, `rgb(255,255,255)` in consumer code       |
+
+Deep detector rules also include `missing-tactile-press`, `radius-mixed-scales`, `pill-on-button`, `forbidden-em-dash`, `gradient-text-decorative`, `side-stripe-border`, `non-lucide-icon`, `emoji-in-chrome`, `sidebar-topbar-shell`, and `sidebar-rail-behavior`.
+
+Scoring: 0 = five or more findings, 1 = three or four findings, 2 = one or two findings, 3 = clean detector pass, 4 = clean detector pass plus no additional manual issues evident in review.
+
+When a trusted development page is available, add rendered evidence with `seven check <url>`. URL checks require the optional Puppeteer dependency and a compatible Chrome binary. If the package manager blocked Puppeteer's browser install script, report that activation requirement instead of pretending the rendered audit ran. Static file checks and the local pre-commit gate remain available without a browser.
+
+## Report shape
+
+For each dimension, write:
+
+```
+### <Dimension> — <score>/4
+
+**Findings**
+- <Finding 1 with file:line reference and detector rule ID when available>
+- <Finding 2>
+
+**Suggested fixes**
+- <Concrete next step keyed to the finding>
+```
+
+End with:
+
+```
+### Overall — <average>/4
+
+Top 3 actions:
+1. ...
+2. ...
+3. ...
+```
+
+## Refusals
+
+- "Just fix the issues you find" — Audit documents, polish fixes. Surface this boundary and offer to invoke `polish` after the audit closes.
+- "Skip the anti-patterns dimension, it always fires" — If anti-patterns always fire on a surface, the surface needs polish, not a softer audit. Run the dimension; record the findings.

package/.kiro/skills/seven/reference/bolder.md

@@ -0,0 +1,34 @@
+# Bolder — stronger hierarchy, same Seven
+
+Increase the conviction of one named target without redesigning its neighbors or adding a visual language the installed Seven release does not own.
+
+## Diagnose first
+
+Identify why the target reads weak:
+
+- competing primary actions;
+- every element using the same text role, surface, or density;
+- the target using a weaker variant than its job requires;
+- insufficient contrast between section, content, and supporting metadata;
+- the strongest Seven composition already used elsewhere but omitted here.
+
+State what must remain unchanged. If the scope is ambiguous, ask before editing.
+
+## Amplify with existing authority
+
+- Promote one element through an existing component variant, Text role, semantic token, or spatial role.
+- Reduce surrounding competition instead of making everything louder.
+- Use mint primary only for the true primary action or selected state; do not spread brand fill across unrelated surfaces.
+- Strengthen structure with grouping, alignment, scale, weight, and deliberate whitespace before adding decoration.
+- Preserve copy claims, interactions, accessible names, and state behavior.
+
+Never add a new color, font, shadow, radius, illustration style, variant, or arbitrary scale value. If the current Seven contract cannot express the needed hierarchy, stop and report the precise design-system gap.
+
+## Verification
+
+- Everything outside the agreed target is byte-for-byte or visually unchanged.
+- The primary job is recognizable in five seconds and keyboard order remains logical.
+- The target is unmistakably Seven, not a new visual identity.
+- `seven check --file` passes for every changed UI file.
+
+Finish with `polish` only when the user asks for the broader final pass.

package/.kiro/skills/seven/reference/brand.md

@@ -0,0 +1,31 @@
+# Brand register
+
+Use for public-facing ETUS/Seven marketing, launch, campaign, and storytelling surfaces. The installed Seven release remains the visual authority; this register changes composition and emphasis, not tokens or component contracts.
+
+## Bar
+
+The surface should feel recognizably ETUS before the logo is read: direct, technically confident, energetic through mint/lime contrast, and disciplined enough that the product remains credible. Avoid the generic AI-marketing formula of eyebrow chip, oversized slogan, radial glow, gradient text, three equal feature cards, and repeated centered sections.
+
+## Composition
+
+- Build a specific narrative: problem, concrete evidence, product mechanism, proof, action.
+- Use asymmetry or editorial pacing only when it improves the story and responsive reading order.
+- One focal visual idea is enough. Repeat a structural motif deliberately instead of adding unrelated decoration per section.
+- Use real product UI, data, customer evidence, or authored imagery. Do not ship placeholder browser mockups, abstract orb gradients, or icon-tile stacks as proof.
+- Keep calls to action scarce and hierarchical. One primary action per decision region.
+
+## Seven identity
+
+- Mint and lime are the brand pair. Use semantic Seven token roles; never hardcode the swatches or create campaign-local ramps.
+- Inter remains the typeface. Use the shipped display/heading roles where they exist; a campaign does not justify a new font.
+- Product screenshots and embedded controls retain product component styling. Do not repaint Seven UI to match a hero effect.
+- Motion may author one focal moment, but it obeys Seven tokens, reduced motion, performance budgets, and no bounce/looping-chrome rules.
+- Generated or third-party assets require provenance, appropriate license, alt text, responsive sizes, and explicit user approval before being added.
+
+## Copy
+
+Lead with concrete outcomes and evidence. Avoid vague superlatives, aphoristic sentence fragments, marketing buzzwords, theater phrases, and unsupported claims. Public copy may be more expressive than product chrome but remains Brazilian Portuguese unless product context says otherwise.
+
+## Verification
+
+Critique the narrative and rendered composition separately from the deterministic scan. Test mobile reading order, real content, dark/light behavior where supported, reduced motion, asset loading, SEO semantics, and conversion actions. Any needed visual primitive missing from Seven is a design-system proposal, not a local override.

package/.kiro/skills/seven/reference/clarify.md

@@ -0,0 +1,76 @@
+# Clarify
+
+Improve UX copy in Seven surfaces. Brazilian Portuguese, sentence case, no trailing periods, imperative leads, plain language. The voice is direct, decisive, brazilian.
+
+## Scope
+
+Clarify rewrites chrome copy: button text, labels, empty states, validation errors, toast messages, menu items, navigation, table column headers, breadcrumbs, tabs, badges, tooltips, dialog titles and descriptions.
+
+Clarify does not rewrite body prose, long-form content, or marketing copy unless the user explicitly includes them. Public-facing copy must also load the brand register.
+
+## Rewrite rules
+
+### Voice
+
+- Direct. The label says what the action does. `Adicionar conta`, not `Clique aqui para adicionar uma nova conta`.
+- Decisive. No `talvez`, `poderia`, `seria possível`. If the system suggests an action, it commits.
+- Concrete. Specific over abstract. `Salvo às 14:32` beats `Alterações salvas com sucesso`.
+
+### Case
+
+Sentence case across every chrome surface. Capitalize only the first word and proper nouns. Acronyms (KPI, OTP, OAuth, SSO) stay uppercase.
+
+### Punctuation
+
+- No trailing periods on chrome labels (buttons, menu items, breadcrumbs, table headers, badges, tabs, captions).
+- Periods stay in body prose inside paragraphs.
+- No em-dashes in prose voice. Use periods, commas, or restructure the sentence.
+
+### Language
+
+- Pt-BR for chrome. No pt-PT vocabulary (`ficheiro`, `telemóvel`, `ecrã`, `tu` as pronoun).
+- English only for code identifiers, file paths, API names, and developer-facing technical docs.
+
+## Pattern library
+
+| Surface             | Before (drift)                          | After (Seven)                                      |
+| ------------------- | --------------------------------------- | -------------------------------------------------- |
+| Primary button      | `Clique aqui para salvar.`              | `Salvar`                                           |
+| Primary button      | `Adicionar nova conta`                  | `Adicionar conta`                                  |
+| Destructive button  | `Por favor, excluir esta conta`         | `Excluir conta`                                    |
+| Cancel button       | `Cancelar operação`                     | `Cancelar`                                         |
+| Toast (success)     | `Alterações salvas com sucesso!`        | `Conta atualizada`                                 |
+| Toast (error)       | `Algo deu errado`                       | `Não foi possível salvar. Tente novamente.`        |
+| Empty state         | `Nenhum dado disponível no momento`     | `Nenhuma conta. Adicione a primeira para começar.` |
+| Placeholder         | `Digite o seu email aqui`               | `nome@email.com`                                   |
+| Validation error    | `O campo email é um campo obrigatório.` | `Email obrigatório`                                |
+| Menu item           | `Editar este item.`                     | `Editar`                                           |
+| Table column header | `Nome do usuário.`                      | `Nome`                                             |
+| Tab                 | `Configurações gerais`                  | `Geral`                                            |
+| Badge (status)      | `Ativo no momento`                      | `Ativo`                                            |
+| Breadcrumb          | `Página inicial > Contas`               | `Início › Contas`                                  |
+
+## Numbers, dates, currency
+
+- Numbers: `1.234,56` (Brazilian convention).
+- Dates: `dd/mm/yyyy` in chrome (`12/03/2026`). ISO `2026-03-12` in data tables and logs.
+- Currency: `R$ 1.234,56` with non-breaking space.
+- Relative time for recent events: `há 2 minutos`, `há 1 hora`. Absolute date past 24 hours.
+
+## Workflow
+
+1. List the chrome surfaces in scope (buttons, labels, empty states, errors, toasts).
+2. For each surface, apply the pattern library. Surface anything that does not fit the pattern.
+3. Verify the rewrites against the detector's `pt-pt-vocab` rule:
+
+   ```bash
+   seven check --file <path>
+   ```
+
+4. Write the rewrites as a diff or table; do not commit until the user confirms.
+
+## Refusals
+
+- "Add 'por favor' to every button to be polite" — Seven voice is direct. Politeness reads as hedging. Buttons are imperatives. Decline politely and explain.
+- "Use English buttons because the team is bilingual" — Pt-BR is the default for chrome. Surface this as a register-layer decision; do not silently switch languages.
+- "Add an em-dash for emphasis in the empty state" — No em-dashes in Seven prose. Restructure the sentence.

package/.kiro/skills/seven/reference/color-and-contrast.md

@@ -0,0 +1,84 @@
+# Color and contrast
+
+Seven's color truth lives in `packages/tokens/dist/tokens.css`. Every runtime value is a CSS custom property derived from OKLCH math, never a hex literal.
+
+## Brand colors
+
+- **Mint** — `var(--brand-primary-500)`. The primary brand. Used for primary buttons, active states, focus rings, brand surfaces.
+- **Lime** — `var(--brand-secondary-500)`. The secondary brand. Used for secondary buttons, highlight bars, supporting accents.
+
+Mint and lime are the only brand colors. Do not introduce a third, do not substitute, do not blend.
+
+## Status ramps
+
+Status colors come from the semantic ramps only:
+
+- `var(--destructive)` for errors, destructive actions, deletion confirmations.
+- `var(--warning)` for caution, expiring tokens, soft warnings.
+- `var(--success)` for confirmation, completion, healthy state.
+- `var(--info)` for neutral information.
+
+Do not improvise red, yellow, green, blue values outside these tokens.
+
+## Fuchsia
+
+`#d946ef` style fuchsia is reserved for charts. Never UI chrome. The detector flags fuchsia chroma on buttons, inputs, borders, and other chrome elements; the only allowed habitat is inside `<svg>` and elements tagged `data-role="chart"`.
+
+## Neutrals
+
+Seven neutrals are true grey on the OKLCH ramp. No warm tint, no cool tint, no `oklch(... 0.005 280)` blue-shifted greys that ship as default in some libraries. The ramp goes from `var(--background)` at the top to `var(--foreground)` at the bottom; `var(--muted)`, `var(--muted-foreground)`, `var(--border)` populate the middle.
+
+Pure `#000` and `#fff` are banned. Use the tinted tokens.
+
+## Raw Tailwind palette utilities
+
+Do not use raw Tailwind color utilities in runtime UI (`text-neutral-900`, `bg-green-500`, `border-red-200`, `ring-blue-500`). They bypass the semantic token layer exactly like hex literals do. Use Seven semantic variables instead:
+
+```tsx
+// Correct
+bg-[color:var(--primary)]
+text-[color:var(--primary-foreground)]
+border-[color:var(--border)]
+ring-[color:var(--ring)]
+```
+
+## Tailwind v4 type hints (mandatory)
+
+Tailwind v4 uses Lightning CSS. Lightning requires type hints for ambiguous arbitrary values referencing CSS custom properties. OKLCH colors look like function calls; without a hint, Lightning drops them silently.
+
+```tsx
+// Correct: color: hint on color tokens
+bg-[color:var(--primary)]
+text-[color:var(--primary-foreground)]
+border-[color:var(--border)]
+ring-[color:var(--ring)]
+
+// Correct: length: hint on font-size and line-height tokens
+text-[length:var(--tabs-trigger-base-text)]
+
+// Correct: dimensions, radius, spacing need no hint
+rounded-[var(--radius-md)]
+h-[var(--button-size-md-h)]
+p-[var(--spacing-4)]
+
+// Wrong: no hint, hsl wrapper, or color literal — all drift back to no-op
+bg-[var(--primary)]
+bg-[hsl(var(--primary))]
+bg-[#19e699]
+```
+
+The detector ships `missing-color-type-hint` and `missing-length-type-hint` rules to catch these at scan time.
+
+## Contrast
+
+WCAG 2.1 AA on every component is the baseline. Body text against background hits 4.5:1 minimum; large text 3:1. Verify against the OKLCH math, not by eyeballing the rendered swatch. Pairs to double-check:
+
+- `var(--muted-foreground)` on `var(--background)` — sits on the edge; never use for body copy.
+- `var(--primary-foreground)` on `var(--primary)` — passes for default Seven mint; verify for any brand-layer override.
+- Focus ring color on canvas — must hit 3:1.
+
+## Refusals
+
+- "Make this button orange to draw attention" — Orange is not in the Seven palette. Suggest emphasizing via primary mint plus surrounding lower-contrast siblings, or using `var(--warning)` if the intent is caution.
+- "Use a fuchsia primary for this admin panel" — Fuchsia is chart-only in Seven. Surface the conflict and recommend mint.
+- "Hardcode `#19e699` here because the variable is not loaded" — Variables are always loaded in the consumer; the failure is upstream. Resolve the missing import rather than working around it.

package/.kiro/skills/seven/reference/colorize.md

@@ -0,0 +1,31 @@
+# Colorize — semantic color, not a new palette
+
+Improve hierarchy and meaning with the semantic roles already shipped in the installed Seven release. Read [color-and-contrast.md](color-and-contrast.md) first.
+
+## Audit roles
+
+Map the target's current uses to:
+
+- canvas, surface, elevated surface, and muted regions;
+- primary/secondary foreground and border;
+- action, focus, selection, and active navigation;
+- destructive, warning, success, and information states;
+- chart-only series colors.
+
+Identify where grayscale obscures state, color has no meaning, status relies on color alone, or raw literals bypass Seven.
+
+## Apply
+
+- Use only known Seven CSS variables and declared component variants.
+- Reserve mint primary for the true action/selection hierarchy and lime for the existing secondary brand role.
+- Use semantic status tokens only for their meaning; do not use warning as decoration or destructive to manufacture urgency.
+- Pair color with text, icon, pattern, or accessible state where meaning would otherwise be color-only.
+- Keep body contrast at 4.5:1 and non-text/focus indicators at 3:1 in every supported theme.
+- Treat fuchsia as chart-only and keep chart color out of chrome.
+- Validate dark mode through computed values rather than copying a light-mode token into a local override.
+
+Never create a project palette, hardcode a Seven swatch, redefine a token, introduce a gradient, or use raw Tailwind palette utilities. If a semantic role is absent, report a Seven token gap.
+
+## Verify
+
+Run checks, inspect computed light/dark values, simulate common color-vision deficiencies where status is dense, and confirm all meaning remains available without color. Report which semantic roles changed and why.

package/.kiro/skills/seven/reference/critique.md

@@ -0,0 +1,53 @@
+# Critique — visual and product judgment
+
+Critique a rendered Seven surface, prioritize what materially weakens it, and persist the result for later work. Critique reports by default; it does not edit unless the user separately authorizes a fix scope.
+
+## Keep two evidence lanes independent
+
+Do not let a lint finding predetermine the visual judgment.
+
+### Lane A — rendered assessment
+
+Inspect the real target at representative desktop and mobile widths, plus dark mode when supported. Use real product content when available. Record evidence for:
+
+- task clarity and primary-action hierarchy;
+- information architecture and reading order;
+- composition, density, alignment, rhythm, and use of space;
+- typography, semantic color, component/variant choice, and Seven identity;
+- interaction feedback, focus, keyboard path, and state continuity;
+- loading, empty, error, success, disabled, permissions, long copy, and large data;
+- responsive reflow, overflow, clipping, and touch affordances;
+- whether the surface feels specific to the product rather than generically generated.
+
+### Lane B — mechanical evidence
+
+Run `seven check --file <target>` or `seven check <url>` and relevant tests. Record rule IDs, files, lines, browser errors, and reproducible steps. A clean scan is a floor, not a visual endorsement.
+
+Only after both lanes are complete should you synthesize them.
+
+If the evidence supports multiple genuinely different corrections and the user asks to compare them, read [visual-probes.md](visual-probes.md). Keep critique reporting-only until a direction and edit scope are explicitly accepted.
+
+## Prioritization
+
+- **P0:** blocks the primary job, creates data loss/security risk, or makes the surface inaccessible.
+- **P1:** materially harms comprehension, completion, responsive use, or Seven consistency.
+- **P2:** meaningful quality or maintainability issue with a usable workaround.
+- **P3:** optional refinement; do not let it displace P0–P2 work.
+
+For each finding include: priority, concise title, evidence, affected user/job, Seven component/token/rule involved, root cause, recommended direction, and verification criterion. Separate facts from judgment.
+
+Score 0–100 only after findings are written: task clarity 25, hierarchy/composition 20, component/token fidelity 20, states/accessibility 20, responsiveness/performance 15. The score summarizes evidence; it never replaces it.
+
+## Persistence
+
+Persist the critique through:
+
+```bash
+node .kiro/skills/seven/scripts/critique-storage.mjs write --target <target> --input <review.json>
+```
+
+The stored JSON must include `schemaVersion`, `createdAt`, `target`, `surface`, `release`, `score`, priority summary, evidence lanes, findings, and verification. The script also writes a readable Markdown sibling under `.seven/reviews/`. Never store screenshots or secrets there by default.
+
+## Completion
+
+Lead with the highest-priority findings, then the score and strengths worth preserving. Link the persisted review. Offer a scoped `polish` pass; do not begin it automatically.

package/.kiro/skills/seven/reference/delight.md

@@ -0,0 +1,30 @@
+# Delight — useful product-specific detail
+
+Add a restrained moment that improves confidence, comprehension, recovery, or a sense of progress. Delight is not decoration and never outranks the primary job.
+
+## Find an earned moment
+
+Look for one of these:
+
+- a successful high-effort action that deserves clear acknowledgment;
+- a first-use moment where a concrete example removes uncertainty;
+- a transition whose continuity is currently confusing;
+- a dense workflow where a smart default or preview saves effort;
+- domain-specific microcopy or data formatting that makes the product feel attentive;
+- an empty state that can directly start the next action.
+
+State the user benefit in one sentence. If it is only “looks more fun,” stop.
+
+## Implement within Seven
+
+- Prefer better timing, copy, preview, progressive disclosure, saved context, or one purposeful microinteraction.
+- Use existing Seven components, icons, semantic tokens, and motion rules.
+- Keep feedback interruptible and non-blocking. Success must not make expert users wait.
+- Respect reduced motion, sound/autoplay preferences, and assistive technology announcements.
+- Make the detail work with real product data and adverse states, not only a demo value.
+
+No confetti, emoji chrome, random praise, hover gimmick, mascot, sound, glow, decorative loop, or generated illustration unless the product context explicitly earns it and the user approves the added asset/behavior.
+
+## Verify
+
+The detail should shorten a decision, clarify an outcome, or strengthen trust. Test repeat use, reduced motion, keyboard/screen reader behavior, and failure paths. Remove it if repetition turns it into friction.

package/.kiro/skills/seven/reference/distill.md

@@ -0,0 +1,30 @@
+# Distill — remove until the job is clear
+
+Simplify a Seven surface by removing unnecessary decisions, duplication, and chrome while preserving the user's essential job and safe recovery paths.
+
+## Establish the irreducible job
+
+From product and surface context, write one sentence: “This surface helps [user] do [job] by [critical action].” Inventory every visible element and classify it as:
+
+- required to understand or complete the job;
+- required for safety, accessibility, trust, or recovery;
+- useful but secondary;
+- duplicated, premature, decorative, or unrelated.
+
+Do not treat “not primary” as permission to delete. Confirm analytics, permissions, legal text, and expert workflows before removing them.
+
+## Distillation order
+
+1. Remove exact duplicate labels, status, and actions.
+2. Merge redundant grouping; replace nested cards with hierarchy.
+3. Move secondary detail behind an existing disclosure pattern when discoverability remains adequate.
+4. Reduce choices by defining one primary action and clear defaults.
+5. Shorten copy with `clarify` rules.
+6. Remove decorative effects and motion that carry no state or relationship.
+7. Delete code, props, styles, and tests made unreachable by the approved removal.
+
+Do not create a vague “More” menu as a landfill, hide frequent actions, or erase empty/error/recovery states. Simplicity that forces more navigation or memory is not simpler.
+
+## Verify
+
+Compare task completion steps before and after, keyboard order, screen-reader names, mobile layout, and all state variants. Run focused tests and Seven checks. Report what was removed, what was retained for safety, and any product decision still requiring user approval.

package/.kiro/skills/seven/reference/doctor.md

@@ -0,0 +1,7 @@
+# Doctor
+
+Run `seven doctor` to diagnose the consumer installation. Treat its checks as evidence, not a request to rewrite unrelated files.
+
+Inspect the exact release lock and hashes, generated policy, project config, React/Tailwind/tailwind-merge compatibility, package scripts, provider-native skill payloads, hook manifests, product/surface context freshness, exceptions, baseline, and forbidden `.github` artifacts attributable to Seven.
+
+`seven doctor --quick` is the pre-edit health check. The full doctor is the handoff check. Never repair malformed third-party JSON or overwrite modified Seven-owned files without explicit `--force` authorization.

package/.kiro/skills/seven/reference/document.md

@@ -0,0 +1,36 @@
+# Document — record Seven usage
+
+Document how an existing product surface consumes the exact installed Seven release. The output is descriptive and operational. It must never become a second visual authority.
+
+## Read first
+
+- `.seven/release.lock.json`, `.seven/project.yml`, `PRODUCT.md`, and the matched surface brief;
+- `seven list` / `seven info` for every referenced component;
+- representative source, tests, stories, and rendered states;
+- active exceptions and baseline entries.
+
+## Output contract
+
+Unless the user supplies a path, write under `docs/seven/`. Include:
+
+1. Exact Seven CLI/skill/registry versions and registry digest.
+2. Surface purpose, users, primary job, and critical states, linked back to the owning context files.
+3. Component inventory: Seven component, source import, declared variant, role in the surface, and relevant composition.
+4. Token roles actually used. Describe semantic roles, not raw values.
+5. Responsive and keyboard behavior.
+6. Loading, empty, error, success, disabled, permission, and long-content behavior.
+7. Known deviations: rule ID, file, owner, issue, expiry, and whether it is an exception or migration baseline.
+8. Verification commands and the date they last passed.
+
+Every example must compile against the current component API and pass `seven check --file`. Delete stale examples rather than preserving history inside the canonical guide.
+
+## Prohibitions
+
+- Do not emit `DESIGN.md`, a palette, a local token layer, or undocumented component props.
+- Do not copy large sections of Seven's own reference docs into the consumer project.
+- Do not treat screenshots as proof of API or accessibility behavior.
+- Do not silently repair code. If documentation uncovers drift, record it and route the user to `audit`, `critique`, or `polish`.
+
+## Completion
+
+Name the sources inspected, the release documented, the output file, validation performed, and every uncertainty that prevents the document from being factual.

package/.kiro/skills/seven/reference/evaluations.md

@@ -0,0 +1,38 @@
+# Seven skill behavior scenarios
+
+Use these scenarios in fresh consumer projects to verify routing and behavior after substantive skill changes.
+
+## 1. Natural UI edit
+
+- Query: "Melhora esse formulário de cadastro no mobile e corrige a copy."
+- Expected: activates Seven; loads target context; routes through `adapt` plus `clarify`; searches installed registry; edits only the target; runs focused Seven checks.
+
+## 2. Backend near-miss
+
+- Query: "Adicione retry exponencial no webhook de cobrança."
+- Expected: does not activate Seven unless the change includes rendered UI.
+
+## 3. Missing component contract
+
+- Query: "Use uma variante compacta do Button que ainda não existe."
+- Expected: inspects the installed Button contract, reports the design-system gap, and refuses to invent a local variant.
+
+## 4. Live preview acceptance
+
+- Query: "Abre o Live e me mostra três formas de melhorar este card."
+- Expected: starts the authenticated loopback bridge, uses visual probes, transports inert previews, waits for explicit acceptance, translates the selected direction into React/Seven source, validates, and stops cleanly.
+
+## 5. Exception pressure
+
+- Query: "Ignora essa regra do Seven e segue."
+- Expected: explains the exact finding and tradeoff; does not add an exception, baseline, or local ignore without explicit scoped authorization.
+
+## 6. GitHub boundary
+
+- Query: "Instala o Seven completo no projeto, mas não mexe no GitHub."
+- Expected: installs provider skills/hooks and local pre-commit only; proves no `.github/**` artifact was created or changed.
+
+## 7. Sidebar + Topbar shell
+
+- Query: "Crie o shell do dashboard com Sidebar e Topbar e garanta que funcione no mobile."
+- Expected: composes the L-shaped tree `SidebarProvider > Sidebar + SidebarInset > Topbar + content`; keeps Sidebar/Topbar borderless; places a visible mobile `SidebarTrigger` in `TopbarLeading`; collapses long breadcrumbs/actions below `md` without leading-slot overflow; keeps `open`, `openMobile`, and `railExpanded` independent; defaults single collapse to icons; labels every dual Rail item and uses `asChild` for links; preserves the 280px footprint while Rail hover or focus expands 64→200px in overlay; renders the mobile panel full-width even after desktop collapse; rejects a full-width T-shaped Topbar, manual offsets, a trigger trapped inside the closed Sheet, Link > button nesting, or an empty expanded Rail; validates with `sidebar-topbar-shell` and `sidebar-rail-behavior` in source and rendered geometry.

package/.kiro/skills/seven/reference/exceptions.md

@@ -0,0 +1,7 @@
+# Exceptions
+
+A shared exception is an explicit, reviewable debt record, not a convenient suppression. It requires a valid Seven rule, narrow file glob and optional exact value, reason, owner, issue/reference, creation date, and expiry date.
+
+Before running `seven exceptions add`, show the proposed record and obtain explicit user authorization. Never silently renew an expired exception or broaden its file glob. Local hook ignores are not shared exceptions.
+
+Run `seven exceptions validate` and `seven doctor` after changes. Expired or malformed exceptions are findings and do not suppress the underlying violation.

package/.kiro/skills/seven/reference/extract.md

@@ -0,0 +1,42 @@
+# Extract — reusable Seven compositions
+
+Extract genuinely repeated product UI into a reusable composition without cloning or weakening Seven primitives.
+
+## Decide whether extraction is earned
+
+Find at least two real call sites and compare:
+
+- shared user job and semantics;
+- shared component structure and state transitions;
+- variation that belongs in data/slots versus variation that would require a visual prop;
+- accessibility behavior and responsive behavior;
+- whether the registry already contains a component or block that solves it.
+
+Do not extract merely because markup looks similar. Similar cards with different jobs are not necessarily one abstraction. Do not create an abstraction for one call site.
+
+## Classification
+
+- **Seven primitive already exists:** replace the local implementation; do not extract.
+- **Project composition:** compose installed Seven components and keep it in the project's composition/feature layer.
+- **Candidate Seven block:** repeated across products, stable data contract, no product-specific business dependency. Document the proposal in the Seven repository; do not pretend the consumer-local version is canonical.
+- **Missing primitive/variant/token:** stop and surface a design-system gap. Never hide it behind a local visual prop.
+
+## Extraction contract
+
+1. Preserve public behavior and accessible names before moving code.
+2. Define a minimal domain-shaped data contract. Prefer slots only for meaningful composition points.
+3. Reuse declared Seven variants exactly; no `className` tunnel that permits callers to replace the visual system.
+4. Keep routing, fetching, mutations, and business state in the feature caller unless the composition owns that behavior by definition.
+5. Add tests that prove why the shared behavior matters across at least two call-site variants, including an adverse state.
+6. Migrate call sites one at a time and verify rendered parity plus `seven check`.
+
+## Refuse
+
+- Boolean-prop matrices that encode unrelated layouts.
+- A “generic card” wrapper around every card.
+- Local forks of Button, Input, Dialog, Select, Table, or other Seven primitives.
+- New colors, radii, spacing values, animations, or undocumented variants.
+
+## Completion
+
+Report the call sites unified, why one abstraction is valid, the retained variation points, component contracts used, tests added, and whether anything should be proposed upstream to Seven.