flyee 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +21 -0
- package/README.md +134 -0
- package/bin/install.js +357 -0
- package/bridge/bridge.py +1780 -0
- package/bridge/local_tracker.py +722 -0
- package/core/agents/backend-specialist.md +266 -0
- package/core/agents/code-archaeologist.md +106 -0
- package/core/agents/database-architect.md +226 -0
- package/core/agents/debugger.md +225 -0
- package/core/agents/devops-engineer.md +323 -0
- package/core/agents/documentation-writer.md +104 -0
- package/core/agents/explorer-agent.md +73 -0
- package/core/agents/frontend-specialist.md +743 -0
- package/core/agents/game-developer.md +162 -0
- package/core/agents/mobile-developer.md +377 -0
- package/core/agents/orchestrator.md +416 -0
- package/core/agents/penetration-tester.md +188 -0
- package/core/agents/performance-optimizer.md +187 -0
- package/core/agents/product-manager.md +112 -0
- package/core/agents/product-owner.md +95 -0
- package/core/agents/project-planner.md +470 -0
- package/core/agents/qa-automation-engineer.md +103 -0
- package/core/agents/security-auditor.md +170 -0
- package/core/agents/seo-specialist.md +111 -0
- package/core/agents/stitch-designer.md +190 -0
- package/core/agents/tdd-reviewer.md +282 -0
- package/core/agents/test-engineer.md +158 -0
- package/core/scripts/auto_preview.py +148 -0
- package/core/scripts/checklist.py +243 -0
- package/core/scripts/cost_report.py +149 -0
- package/core/scripts/doc-sync-check.py +461 -0
- package/core/scripts/parse_user_stories.py +79 -0
- package/core/scripts/prepare_notion_updates.py +172 -0
- package/core/scripts/print_create_payload.py +18 -0
- package/core/scripts/session_manager.py +120 -0
- package/core/scripts/task_complete.py +127 -0
- package/core/scripts/verify_all.py +327 -0
- package/core/skills/analytics-strategy/SKILL.md +128 -0
- package/core/skills/api-patterns/SKILL.md +81 -0
- package/core/skills/api-patterns/api-style.md +42 -0
- package/core/skills/api-patterns/auth.md +24 -0
- package/core/skills/api-patterns/documentation.md +26 -0
- package/core/skills/api-patterns/graphql.md +41 -0
- package/core/skills/api-patterns/rate-limiting.md +31 -0
- package/core/skills/api-patterns/response.md +37 -0
- package/core/skills/api-patterns/rest.md +40 -0
- package/core/skills/api-patterns/scripts/api_validator.py +211 -0
- package/core/skills/api-patterns/security-testing.md +122 -0
- package/core/skills/api-patterns/trpc.md +41 -0
- package/core/skills/api-patterns/versioning.md +22 -0
- package/core/skills/app-builder/SKILL.md +75 -0
- package/core/skills/app-builder/agent-coordination.md +71 -0
- package/core/skills/app-builder/feature-building.md +53 -0
- package/core/skills/app-builder/project-detection.md +34 -0
- package/core/skills/app-builder/scaffolding.md +118 -0
- package/core/skills/app-builder/tech-stack.md +40 -0
- package/core/skills/app-builder/templates/SKILL.md +39 -0
- package/core/skills/app-builder/templates/astro-static/TEMPLATE.md +76 -0
- package/core/skills/app-builder/templates/chrome-extension/TEMPLATE.md +92 -0
- package/core/skills/app-builder/templates/cli-tool/TEMPLATE.md +88 -0
- package/core/skills/app-builder/templates/electron-desktop/TEMPLATE.md +88 -0
- package/core/skills/app-builder/templates/express-api/TEMPLATE.md +83 -0
- package/core/skills/app-builder/templates/flutter-app/TEMPLATE.md +90 -0
- package/core/skills/app-builder/templates/monorepo-turborepo/TEMPLATE.md +90 -0
- package/core/skills/app-builder/templates/nextjs-fullstack/TEMPLATE.md +82 -0
- package/core/skills/app-builder/templates/nextjs-saas/TEMPLATE.md +100 -0
- package/core/skills/app-builder/templates/nextjs-static/TEMPLATE.md +106 -0
- package/core/skills/app-builder/templates/nuxt-app/TEMPLATE.md +101 -0
- package/core/skills/app-builder/templates/python-fastapi/TEMPLATE.md +83 -0
- package/core/skills/app-builder/templates/react-native-app/TEMPLATE.md +93 -0
- package/core/skills/architecture/SKILL.md +55 -0
- package/core/skills/architecture/context-discovery.md +43 -0
- package/core/skills/architecture/examples.md +94 -0
- package/core/skills/architecture/pattern-selection.md +68 -0
- package/core/skills/architecture/patterns-reference.md +50 -0
- package/core/skills/architecture/trade-off-analysis.md +77 -0
- package/core/skills/atomic-design/SKILL.md +282 -0
- package/core/skills/atomic-design/references/classification-guide.md +132 -0
- package/core/skills/atomic-design/references/quality-checklist.md +60 -0
- package/core/skills/atomic-design/references/stacks/stack-blade.md +254 -0
- package/core/skills/atomic-design/references/stacks/stack-nextjs.md +272 -0
- package/core/skills/atomic-design/references/stacks/stack-react.md +239 -0
- package/core/skills/atomic-design/references/stacks/stack-vue.md +224 -0
- package/core/skills/bash-linux/SKILL.md +199 -0
- package/core/skills/behavioral-modes/SKILL.md +242 -0
- package/core/skills/brainstorming/SKILL.md +163 -0
- package/core/skills/brainstorming/dynamic-questioning.md +373 -0
- package/core/skills/checkpointing-patterns/SKILL.md +163 -0
- package/core/skills/clean-code/SKILL.md +201 -0
- package/core/skills/code-review-checklist/SKILL.md +109 -0
- package/core/skills/code-truth-validation/SKILL.md +149 -0
- package/core/skills/component-library-discovery/SKILL.md +154 -0
- package/core/skills/content-strategy/SKILL.md +222 -0
- package/core/skills/context-budget/SKILL.md +155 -0
- package/core/skills/context-gathering-patterns/SKILL.md +278 -0
- package/core/skills/cost-tracking/SKILL.md +206 -0
- package/core/skills/database-design/SKILL.md +52 -0
- package/core/skills/database-design/database-selection.md +43 -0
- package/core/skills/database-design/indexing.md +39 -0
- package/core/skills/database-design/migrations.md +48 -0
- package/core/skills/database-design/optimization.md +36 -0
- package/core/skills/database-design/orm-selection.md +30 -0
- package/core/skills/database-design/schema-design.md +56 -0
- package/core/skills/database-design/scripts/schema_validator.py +172 -0
- package/core/skills/deployment-procedures/SKILL.md +295 -0
- package/core/skills/design-md/README.md +34 -0
- package/core/skills/design-md/SKILL.md +172 -0
- package/core/skills/design-md/examples/DESIGN.md +154 -0
- package/core/skills/design-system-enforcement/SKILL.md +339 -0
- package/core/skills/doc.md +177 -0
- package/core/skills/document-registry/SKILL.md +130 -0
- package/core/skills/documentation-publishing/SKILL.md +174 -0
- package/core/skills/documentation-templates/SKILL.md +194 -0
- package/core/skills/enhance-prompt/README.md +34 -0
- package/core/skills/enhance-prompt/SKILL.md +204 -0
- package/core/skills/enhance-prompt/references/KEYWORDS.md +114 -0
- package/core/skills/frontend-design/SKILL.md +430 -0
- package/core/skills/frontend-design/animation-guide.md +331 -0
- package/core/skills/frontend-design/color-system.md +311 -0
- package/core/skills/frontend-design/decision-trees.md +418 -0
- package/core/skills/frontend-design/motion-graphics.md +306 -0
- package/core/skills/frontend-design/scripts/accessibility_checker.py +183 -0
- package/core/skills/frontend-design/scripts/ux_audit.py +722 -0
- package/core/skills/frontend-design/typography-system.md +345 -0
- package/core/skills/frontend-design/ux-psychology.md +541 -0
- package/core/skills/frontend-design/visual-effects.md +383 -0
- package/core/skills/game-development/2d-games/SKILL.md +119 -0
- package/core/skills/game-development/3d-games/SKILL.md +135 -0
- package/core/skills/game-development/SKILL.md +167 -0
- package/core/skills/game-development/game-art/SKILL.md +185 -0
- package/core/skills/game-development/game-audio/SKILL.md +190 -0
- package/core/skills/game-development/game-design/SKILL.md +129 -0
- package/core/skills/game-development/mobile-games/SKILL.md +108 -0
- package/core/skills/game-development/multiplayer/SKILL.md +132 -0
- package/core/skills/game-development/pc-games/SKILL.md +144 -0
- package/core/skills/game-development/vr-ar/SKILL.md +123 -0
- package/core/skills/game-development/web-games/SKILL.md +150 -0
- package/core/skills/geo-fundamentals/SKILL.md +156 -0
- package/core/skills/geo-fundamentals/scripts/geo_checker.py +289 -0
- package/core/skills/git-workflow/SKILL.md +263 -0
- package/core/skills/history-check-patterns/SKILL.md +125 -0
- package/core/skills/i18n-localization/SKILL.md +154 -0
- package/core/skills/i18n-localization/scripts/i18n_checker.py +241 -0
- package/core/skills/integration-completeness/SKILL.md +219 -0
- package/core/skills/intelligent-routing/SKILL.md +370 -0
- package/core/skills/lint-and-validate/SKILL.md +45 -0
- package/core/skills/lint-and-validate/scripts/lint_runner.py +173 -0
- package/core/skills/lint-and-validate/scripts/type_coverage.py +173 -0
- package/core/skills/local-verification/SKILL.md +195 -0
- package/core/skills/mcp-builder/SKILL.md +176 -0
- package/core/skills/mobile-design/SKILL.md +394 -0
- package/core/skills/mobile-design/decision-trees.md +516 -0
- package/core/skills/mobile-design/mobile-backend.md +491 -0
- package/core/skills/mobile-design/mobile-color-system.md +420 -0
- package/core/skills/mobile-design/mobile-debugging.md +122 -0
- package/core/skills/mobile-design/mobile-design-thinking.md +357 -0
- package/core/skills/mobile-design/mobile-navigation.md +458 -0
- package/core/skills/mobile-design/mobile-performance.md +767 -0
- package/core/skills/mobile-design/mobile-testing.md +356 -0
- package/core/skills/mobile-design/mobile-typography.md +433 -0
- package/core/skills/mobile-design/platform-android.md +666 -0
- package/core/skills/mobile-design/platform-ios.md +561 -0
- package/core/skills/mobile-design/scripts/mobile_audit.py +670 -0
- package/core/skills/mobile-design/touch-psychology.md +537 -0
- package/core/skills/nextjs-react-expert/1-async-eliminating-waterfalls.md +312 -0
- package/core/skills/nextjs-react-expert/2-bundle-bundle-size-optimization.md +240 -0
- package/core/skills/nextjs-react-expert/3-server-server-side-performance.md +490 -0
- package/core/skills/nextjs-react-expert/4-client-client-side-data-fetching.md +264 -0
- package/core/skills/nextjs-react-expert/5-rerender-re-render-optimization.md +581 -0
- package/core/skills/nextjs-react-expert/6-rendering-rendering-performance.md +432 -0
- package/core/skills/nextjs-react-expert/7-js-javascript-performance.md +684 -0
- package/core/skills/nextjs-react-expert/8-advanced-advanced-patterns.md +150 -0
- package/core/skills/nextjs-react-expert/SKILL.md +267 -0
- package/core/skills/nextjs-react-expert/scripts/convert_rules.py +222 -0
- package/core/skills/nextjs-react-expert/scripts/react_performance_checker.py +252 -0
- package/core/skills/nodejs-best-practices/SKILL.md +333 -0
- package/core/skills/notion-task-patterns/SKILL.md +2529 -0
- package/core/skills/page-specifications/SKILL.md +367 -0
- package/core/skills/parallel-agents/SKILL.md +175 -0
- package/core/skills/performance-profiling/SKILL.md +143 -0
- package/core/skills/performance-profiling/scripts/lighthouse_audit.py +76 -0
- package/core/skills/plan-writing/SKILL.md +190 -0
- package/core/skills/powershell-windows/SKILL.md +167 -0
- package/core/skills/project-foundation/SKILL.md +117 -0
- package/core/skills/project-setup/SKILL.md +141 -0
- package/core/skills/project-tracking-patterns/SKILL.md +357 -0
- package/core/skills/project-type-discovery/SKILL.md +239 -0
- package/core/skills/python-patterns/SKILL.md +441 -0
- package/core/skills/qa-test-generation/SKILL.md +156 -0
- package/core/skills/react-components/README.md +36 -0
- package/core/skills/react-components/SKILL.md +47 -0
- package/core/skills/react-components/examples/gold-standard-card.tsx +80 -0
- package/core/skills/react-components/package-lock.json +231 -0
- package/core/skills/react-components/package.json +16 -0
- package/core/skills/react-components/resources/architecture-checklist.md +15 -0
- package/core/skills/react-components/resources/component-template.tsx +37 -0
- package/core/skills/react-components/resources/stitch-api-reference.md +14 -0
- package/core/skills/react-components/resources/style-guide.json +27 -0
- package/core/skills/react-components/scripts/fetch-stitch.sh +30 -0
- package/core/skills/react-components/scripts/validate.js +68 -0
- package/core/skills/red-team-tactics/SKILL.md +199 -0
- package/core/skills/remotion/README.md +105 -0
- package/core/skills/remotion/SKILL.md +393 -0
- package/core/skills/remotion/examples/WalkthroughComposition.tsx +78 -0
- package/core/skills/remotion/examples/screens.json +56 -0
- package/core/skills/remotion/resources/composition-checklist.md +124 -0
- package/core/skills/remotion/resources/screen-slide-template.tsx +123 -0
- package/core/skills/remotion/scripts/download-stitch-asset.sh +38 -0
- package/core/skills/seo-fundamentals/SKILL.md +129 -0
- package/core/skills/seo-fundamentals/scripts/seo_checker.py +219 -0
- package/core/skills/server-management/SKILL.md +161 -0
- package/core/skills/session-resilience/SKILL.md +199 -0
- package/core/skills/shadcn-ui/README.md +248 -0
- package/core/skills/shadcn-ui/SKILL.md +326 -0
- package/core/skills/shadcn-ui/examples/auth-layout.tsx +177 -0
- package/core/skills/shadcn-ui/examples/data-table.tsx +313 -0
- package/core/skills/shadcn-ui/examples/form-pattern.tsx +177 -0
- package/core/skills/shadcn-ui/resources/component-catalog.md +481 -0
- package/core/skills/shadcn-ui/resources/customization-guide.md +516 -0
- package/core/skills/shadcn-ui/resources/migration-guide.md +463 -0
- package/core/skills/shadcn-ui/resources/setup-guide.md +412 -0
- package/core/skills/shadcn-ui/scripts/verify-setup.sh +134 -0
- package/core/skills/state-machine/SKILL.md +264 -0
- package/core/skills/stitch-loop/README.md +54 -0
- package/core/skills/stitch-loop/SKILL.md +203 -0
- package/core/skills/stitch-loop/examples/SITE.md +73 -0
- package/core/skills/stitch-loop/examples/next-prompt.md +25 -0
- package/core/skills/stitch-loop/resources/baton-schema.md +61 -0
- package/core/skills/stitch-loop/resources/site-template.md +104 -0
- package/core/skills/systematic-debugging/SKILL.md +109 -0
- package/core/skills/tailwind-patterns/SKILL.md +284 -0
- package/core/skills/tdd-validation/SKILL.md +243 -0
- package/core/skills/tdd-workflow/SKILL.md +284 -0
- package/core/skills/testing-patterns/SKILL.md +196 -0
- package/core/skills/testing-patterns/scripts/test_runner.py +219 -0
- package/core/skills/ui-ux-discovery/SKILL.md +329 -0
- package/core/skills/ui-validation/SKILL.md +190 -0
- package/core/skills/ui-validation/scripts/ui_antipattern_check.py +317 -0
- package/core/skills/verification-gate/SKILL.md +205 -0
- package/core/skills/vulnerability-scanner/SKILL.md +276 -0
- package/core/skills/vulnerability-scanner/checklists.md +121 -0
- package/core/skills/vulnerability-scanner/scripts/security_scan.py +458 -0
- package/core/skills/web-design-guidelines/SKILL.md +57 -0
- package/core/skills/webapp-testing/SKILL.md +187 -0
- package/core/skills/webapp-testing/scripts/playwright_runner.py +173 -0
- package/core/templates/ARCHITECTURE.template.md +407 -0
- package/core/templates/project-resources.example.json +71 -0
- package/core/workflows/atomic.md +182 -0
- package/core/workflows/brainstorm.md +134 -0
- package/core/workflows/check-task.md +242 -0
- package/core/workflows/copy-collect.md +306 -0
- package/core/workflows/create-agent.md +33 -0
- package/core/workflows/create-skill.md +39 -0
- package/core/workflows/create-workflow.md +33 -0
- package/core/workflows/create.md +92 -0
- package/core/workflows/debug.md +186 -0
- package/core/workflows/demand.md +443 -0
- package/core/workflows/deploy.md +260 -0
- package/core/workflows/discovery.md +267 -0
- package/core/workflows/document.md +272 -0
- package/core/workflows/ds-components.md +296 -0
- package/core/workflows/ds-init.md +58 -0
- package/core/workflows/ds-refactor.md +245 -0
- package/core/workflows/ds-references.md +197 -0
- package/core/workflows/ds-styleguide.md +237 -0
- package/core/workflows/ds-token-diff.md +103 -0
- package/core/workflows/ds-tokens.md +317 -0
- package/core/workflows/ds-validate.md +309 -0
- package/core/workflows/execute.md +483 -0
- package/core/workflows/extract-template.md +278 -0
- package/core/workflows/fix-failed-tests.md +160 -0
- package/core/workflows/init-project.md +386 -0
- package/core/workflows/legacy-project.md +849 -0
- package/core/workflows/log.md +97 -0
- package/core/workflows/new-project.md +610 -0
- package/core/workflows/new-project.md.bak +3292 -0
- package/core/workflows/new-task.md +404 -0
- package/core/workflows/orchestrate.md +237 -0
- package/core/workflows/page-build.md +296 -0
- package/core/workflows/plan.md +89 -0
- package/core/workflows/prd.md +255 -0
- package/core/workflows/preview.md +81 -0
- package/core/workflows/review-page.md +304 -0
- package/core/workflows/status.md +86 -0
- package/core/workflows/stitch.md +226 -0
- package/core/workflows/task-complete.md +473 -0
- package/core/workflows/task-update.md +163 -0
- package/core/workflows/tdd.md +344 -0
- package/core/workflows/test.md +251 -0
- package/core/workflows/ui-ux-pro-max.md +437 -0
- package/core/workflows/ux-mobile-optimize.md +262 -0
- package/core/workflows/ux-mobile-validate.md +297 -0
- package/engine-files/GEMINI.md +69 -0
- package/package.json +47 -0
- package/runtime-adapters/antigravity.js +26 -0
- package/runtime-adapters/claude.js +57 -0
- package/runtime-adapters/codex.js +51 -0
- package/runtime-adapters/copilot.js +51 -0
- package/runtime-adapters/cursor.js +51 -0
- package/runtime-adapters/gemini-cli.js +30 -0
- package/runtime-adapters/opencode.js +51 -0
- package/runtime-adapters/windsurf.js +51 -0
|
@@ -0,0 +1,222 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: content-strategy
|
|
3
|
+
description: Define conteúdo textual (copy) para LPs, páginas internas e metadados SEO antes da implementação
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# Content Strategy Skill
|
|
7
|
+
|
|
8
|
+
> **Objetivo:** Definir TODO o conteúdo textual ANTES de implementar UI.
|
|
9
|
+
> Evita retrabalho e garante consistência de mensagem.
|
|
10
|
+
|
|
11
|
+
---
|
|
12
|
+
|
|
13
|
+
## Quando Usar
|
|
14
|
+
|
|
15
|
+
| Cenário | Usar? |
|
|
16
|
+
|---------|-------|
|
|
17
|
+
| SaaS com Landing Page | ✅ Obrigatório |
|
|
18
|
+
| Site institucional | ✅ Obrigatório |
|
|
19
|
+
| E-commerce | ✅ Obrigatório |
|
|
20
|
+
| Dashboard interno | ⚠️ Opcional (empty states, labels) |
|
|
21
|
+
| API/Backend puro | ❌ Pular |
|
|
22
|
+
|
|
23
|
+
---
|
|
24
|
+
|
|
25
|
+
## Processo (4 Passos)
|
|
26
|
+
|
|
27
|
+
### PASSO 1: Identificar Páginas
|
|
28
|
+
|
|
29
|
+
Listar todas as páginas que precisam de conteúdo:
|
|
30
|
+
|
|
31
|
+
```markdown
|
|
32
|
+
## Páginas Identificadas
|
|
33
|
+
|
|
34
|
+
### Públicas (Marketing) - OBRIGATÓRIAS para SaaS/Produto
|
|
35
|
+
- [ ] Landing Page `/` (Hero, Features, Pricing Preview, FAQ Preview)
|
|
36
|
+
- [ ] Iniciar Projeto `/start` (Wizard Entry - Link padrão YouTube)
|
|
37
|
+
- [ ] Pricing `/pricing` (Detalhado, com FAQs próprias)
|
|
38
|
+
- [ ] Como Funciona `/how-it-works` (Educação rápida, não curso)
|
|
39
|
+
- [ ] Biblioteca/Catálogo `/workflows` ou `/features` (Semi-aberto: público + premium)
|
|
40
|
+
|
|
41
|
+
### Públicas (Opcional)
|
|
42
|
+
- [ ] Blog `/blog` (se estratégia de conteúdo)
|
|
43
|
+
- [ ] Comunidade `/community` (delimitadora, filtra público errado)
|
|
44
|
+
- [ ] Páginas por Vídeo `/videos/{slug}` (CMS, SEO, ponte YouTube→Produto)
|
|
45
|
+
|
|
46
|
+
### Privadas (App)
|
|
47
|
+
- [ ] Dashboard
|
|
48
|
+
- [ ] Onboarding/Wizard
|
|
49
|
+
- [ ] Settings
|
|
50
|
+
- [ ] Profile
|
|
51
|
+
```
|
|
52
|
+
|
|
53
|
+
> [!IMPORTANT]
|
|
54
|
+
> **Para SaaS/Produtos com YouTube:** A página `/start` deve ser o link padrão dos vídeos.
|
|
55
|
+
> As páginas de vídeo (`/videos/{slug}`) devem ser CMS, não hardcoded.
|
|
56
|
+
|
|
57
|
+
---
|
|
58
|
+
|
|
59
|
+
### PASSO 2: Perguntas ao Usuário ⭐
|
|
60
|
+
|
|
61
|
+
> [!CAUTION]
|
|
62
|
+
> **BLOQUEADOR:** Fazer estas perguntas ANTES de definir o copy.
|
|
63
|
+
|
|
64
|
+
#### Perguntas Obrigatórias
|
|
65
|
+
|
|
66
|
+
1. **Tom de Voz:**
|
|
67
|
+
- A) Formal/Corporativo
|
|
68
|
+
- B) Profissional mas acessível
|
|
69
|
+
- C) Casual/Amigável
|
|
70
|
+
- D) Técnico/Developer-focused
|
|
71
|
+
|
|
72
|
+
2. **Público Principal:**
|
|
73
|
+
- Quem é o usuário ideal? (persona)
|
|
74
|
+
- Qual o nível técnico?
|
|
75
|
+
|
|
76
|
+
3. **Proposta de Valor:**
|
|
77
|
+
- Qual o principal benefício em UMA frase?
|
|
78
|
+
- O que diferencia dos concorrentes?
|
|
79
|
+
|
|
80
|
+
4. **Urgência/CTA:**
|
|
81
|
+
- Qual ação principal queremos que o usuário tome?
|
|
82
|
+
- Há senso de urgência? (Early access, Limited spots, etc.)
|
|
83
|
+
|
|
84
|
+
5. **Pricing (se aplicável):**
|
|
85
|
+
- Quantos planos?
|
|
86
|
+
- Qual a estrutura de preços?
|
|
87
|
+
- Há trial/free tier?
|
|
88
|
+
|
|
89
|
+
---
|
|
90
|
+
|
|
91
|
+
### PASSO 3: Gerar Documento
|
|
92
|
+
|
|
93
|
+
Criar `docs/content/CONTENT-STRATEGY-{Projeto}.md` com:
|
|
94
|
+
|
|
95
|
+
```markdown
|
|
96
|
+
# Content Strategy - {Projeto}
|
|
97
|
+
|
|
98
|
+
> **Tom de Voz:** {resposta}
|
|
99
|
+
> **Público:** {resposta}
|
|
100
|
+
> **USP:** {resposta em 1 linha}
|
|
101
|
+
|
|
102
|
+
---
|
|
103
|
+
|
|
104
|
+
## 1. Landing Page
|
|
105
|
+
|
|
106
|
+
### Hero Section
|
|
107
|
+
| Elemento | Conteúdo |
|
|
108
|
+
|----------|----------|
|
|
109
|
+
| **Headline** | {título principal - max 10 palavras} |
|
|
110
|
+
| **Subheadline** | {descrição - max 25 palavras} |
|
|
111
|
+
| **CTA Primário** | {ex: "Começar Grátis"} |
|
|
112
|
+
| **CTA Secundário** | {ex: "Ver Demo"} |
|
|
113
|
+
|
|
114
|
+
### Features (3-6 cards)
|
|
115
|
+
| # | Título | Descrição (max 15 palavras) | Ícone |
|
|
116
|
+
|---|--------|------------------------------|-------|
|
|
117
|
+
| 1 | | | |
|
|
118
|
+
| 2 | | | |
|
|
119
|
+
| 3 | | | |
|
|
120
|
+
|
|
121
|
+
### How it Works (3-5 steps)
|
|
122
|
+
| Step | Título | Descrição |
|
|
123
|
+
|------|--------|-----------|
|
|
124
|
+
| 1 | | |
|
|
125
|
+
| 2 | | |
|
|
126
|
+
| 3 | | |
|
|
127
|
+
|
|
128
|
+
### Social Proof
|
|
129
|
+
- [ ] Testimonials (quantos?)
|
|
130
|
+
- [ ] Logos de clientes
|
|
131
|
+
- [ ] Métricas (ex: "10k+ usuários")
|
|
132
|
+
|
|
133
|
+
### Pricing (se aplicável)
|
|
134
|
+
| Plano | Preço | Features |
|
|
135
|
+
|-------|-------|----------|
|
|
136
|
+
| Free | $0 | |
|
|
137
|
+
| Pro | $ | |
|
|
138
|
+
| Enterprise | $ | |
|
|
139
|
+
|
|
140
|
+
### FAQ (5-10 perguntas)
|
|
141
|
+
| # | Pergunta | Resposta |
|
|
142
|
+
|---|----------|----------|
|
|
143
|
+
| 1 | | |
|
|
144
|
+
| 2 | | |
|
|
145
|
+
|
|
146
|
+
### Footer
|
|
147
|
+
- **Links:** Home, Features, Pricing, FAQ, Contact, Terms, Privacy
|
|
148
|
+
- **Social:** Twitter, LinkedIn, GitHub (quais?)
|
|
149
|
+
|
|
150
|
+
---
|
|
151
|
+
|
|
152
|
+
## 2. Páginas Internas
|
|
153
|
+
|
|
154
|
+
### Empty States
|
|
155
|
+
| Página | Mensagem quando vazio | CTA |
|
|
156
|
+
|--------|----------------------|-----|
|
|
157
|
+
| Dashboard | | |
|
|
158
|
+
| Projects | | |
|
|
159
|
+
|
|
160
|
+
### Labels de Navegação
|
|
161
|
+
| Seção | Label |
|
|
162
|
+
|-------|-------|
|
|
163
|
+
| Home | Dashboard |
|
|
164
|
+
| | |
|
|
165
|
+
|
|
166
|
+
### Onboarding Steps (se aplicável)
|
|
167
|
+
| Step | Título | Instrução | Placeholder |
|
|
168
|
+
|------|--------|-----------|-------------|
|
|
169
|
+
| 1 | | | |
|
|
170
|
+
|
|
171
|
+
---
|
|
172
|
+
|
|
173
|
+
## 3. SEO Metadata
|
|
174
|
+
|
|
175
|
+
| Página | Title (max 60 chars) | Meta Description (max 155 chars) |
|
|
176
|
+
|--------|---------------------|----------------------------------|
|
|
177
|
+
| Home | | |
|
|
178
|
+
| Pricing | | |
|
|
179
|
+
| Dashboard | | |
|
|
180
|
+
```
|
|
181
|
+
|
|
182
|
+
---
|
|
183
|
+
|
|
184
|
+
### PASSO 4: Validar e Aprovar
|
|
185
|
+
|
|
186
|
+
**Gate de Saída:**
|
|
187
|
+
```
|
|
188
|
+
[ ] Documento CONTENT-STRATEGY gerado
|
|
189
|
+
[ ] Hero copy definido (headline, subheadline, CTAs)
|
|
190
|
+
[ ] Features descritas
|
|
191
|
+
[ ] Pricing documentado (se aplicável)
|
|
192
|
+
[ ] FAQ com pelo menos 5 perguntas
|
|
193
|
+
[ ] SEO metadata para páginas principais
|
|
194
|
+
[ ] Conteúdo APROVADO pelo humano
|
|
195
|
+
```
|
|
196
|
+
|
|
197
|
+
> [!CAUTION]
|
|
198
|
+
> **BLOQUEADOR:** Não prosseguir para implementação sem aprovação do Content Strategy.
|
|
199
|
+
|
|
200
|
+
---
|
|
201
|
+
|
|
202
|
+
## Integração com Workflows
|
|
203
|
+
|
|
204
|
+
Esta skill é chamada por:
|
|
205
|
+
- `/new-project` → Phase 2.75
|
|
206
|
+
- `/discovery` → Fase 4.5
|
|
207
|
+
- `/legacy-project` → Phase 5.75
|
|
208
|
+
|
|
209
|
+
**Referência nos workflows:**
|
|
210
|
+
```markdown
|
|
211
|
+
> [!IMPORTANT]
|
|
212
|
+
> **SKILL:** Seguir `content-strategy` para definição de copy e conteúdo.
|
|
213
|
+
```
|
|
214
|
+
|
|
215
|
+
---
|
|
216
|
+
|
|
217
|
+
## Flags Relacionadas
|
|
218
|
+
|
|
219
|
+
| Flag | Efeito |
|
|
220
|
+
|------|--------|
|
|
221
|
+
| `--no-content` | Pula esta fase (para APIs/backends) |
|
|
222
|
+
| `--minimal-content` | Apenas Hero + Features (sem FAQ/Pricing detalhado) |
|
|
@@ -0,0 +1,155 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: context-budget
|
|
3
|
+
description: Enforces that tasks fit within a single AI context window. Prevents scope creep. Task splitting protocol. Absorbed from GSD-2 "iron rule" principle.
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# Context Budget
|
|
7
|
+
|
|
8
|
+
> **Absorbed from:** GSD-2 "Iron Rule" — one task must fit in one context window.
|
|
9
|
+
> **Purpose:** Prevent tasks that are too large to complete in a single session.
|
|
10
|
+
|
|
11
|
+
---
|
|
12
|
+
|
|
13
|
+
## 🎯 THE IRON RULE
|
|
14
|
+
|
|
15
|
+
> **A single task MUST be completable within one context window.**
|
|
16
|
+
> If it can't, it's not a task — it's a phase. SPLIT IT.
|
|
17
|
+
|
|
18
|
+
---
|
|
19
|
+
|
|
20
|
+
## 📏 SIZE LIMITS
|
|
21
|
+
|
|
22
|
+
| Metric | Threshold | Action if exceeded |
|
|
23
|
+
|--------|-----------|-------------------|
|
|
24
|
+
| **Files modified** | > 8 files | ⚠️ Consider splitting |
|
|
25
|
+
| **Files modified** | > 12 files | 🔴 MUST split |
|
|
26
|
+
| **Lines written** | > 400 lines | ⚠️ Consider splitting |
|
|
27
|
+
| **Lines written** | > 600 lines | 🔴 MUST split |
|
|
28
|
+
| **Steps in plan** | > 8 steps | ⚠️ Consider splitting |
|
|
29
|
+
| **Steps in plan** | > 12 steps | 🔴 MUST split |
|
|
30
|
+
| **External dependencies** | > 3 new packages | ⚠️ Consider splitting |
|
|
31
|
+
|
|
32
|
+
---
|
|
33
|
+
|
|
34
|
+
## 🔄 SPLITTING PROTOCOL
|
|
35
|
+
|
|
36
|
+
When a task exceeds limits:
|
|
37
|
+
|
|
38
|
+
```
|
|
39
|
+
1. IDENTIFY natural boundaries:
|
|
40
|
+
- Backend vs Frontend
|
|
41
|
+
- Data layer vs UI layer
|
|
42
|
+
- Create vs Connect vs Style
|
|
43
|
+
- Core logic vs Edge cases
|
|
44
|
+
|
|
45
|
+
2. SPLIT into sub-tasks:
|
|
46
|
+
- Each sub-task has its own PLAN.md
|
|
47
|
+
- Each sub-task has its own must-haves
|
|
48
|
+
- Each sub-task fits in one context window
|
|
49
|
+
|
|
50
|
+
3. DEFINE boundary map:
|
|
51
|
+
- What does sub-task A produce that B needs?
|
|
52
|
+
- What interfaces/types are shared?
|
|
53
|
+
|
|
54
|
+
4. ORDER by dependency:
|
|
55
|
+
- Tasks with no dependencies first
|
|
56
|
+
- Then tasks that consume from completed ones
|
|
57
|
+
```
|
|
58
|
+
|
|
59
|
+
### Example: "Blog Feature" → Split
|
|
60
|
+
|
|
61
|
+
```
|
|
62
|
+
BEFORE (too large):
|
|
63
|
+
T01 — Build complete blog feature
|
|
64
|
+
|
|
65
|
+
AFTER (properly split):
|
|
66
|
+
T01 — Blog API endpoints (backend only)
|
|
67
|
+
Artifacts: api/posts/route.ts, types/post.ts
|
|
68
|
+
T02 — BlogList page skeleton (frontend structure)
|
|
69
|
+
Consumes: types/post.ts from T01
|
|
70
|
+
Artifacts: pages/blog/page.tsx, components/BlogList.tsx
|
|
71
|
+
T03 — BlogCard component (frontend component)
|
|
72
|
+
Consumes: components/BlogList.tsx from T02
|
|
73
|
+
Artifacts: components/BlogCard.tsx, BlogCard.module.css
|
|
74
|
+
T04 — Blog pagination + filters (frontend polish)
|
|
75
|
+
Consumes: BlogList.tsx from T02, API from T01
|
|
76
|
+
Artifacts: hooks/usePosts.ts (add pagination)
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
---
|
|
80
|
+
|
|
81
|
+
## 📋 CONTEXT LOADING STRATEGY
|
|
82
|
+
|
|
83
|
+
When starting a task, load context efficiently:
|
|
84
|
+
|
|
85
|
+
### Priority Pyramid (low cost → high cost)
|
|
86
|
+
|
|
87
|
+
```
|
|
88
|
+
Level 0: FREE — Already known
|
|
89
|
+
└── STATE.md, continue.md
|
|
90
|
+
|
|
91
|
+
Level 1: CHEAP — Frontmatter only
|
|
92
|
+
└── Read first 10 lines of related files (imports, types)
|
|
93
|
+
|
|
94
|
+
Level 2: MODERATE — Key sections
|
|
95
|
+
└── Read specific functions, not entire files
|
|
96
|
+
|
|
97
|
+
Level 3: EXPENSIVE — Full files
|
|
98
|
+
└── Complete file reads — only when necessary
|
|
99
|
+
```
|
|
100
|
+
|
|
101
|
+
### Loading Rules
|
|
102
|
+
|
|
103
|
+
1. **Start with STATE.md** → know where you are
|
|
104
|
+
2. **Read task PLAN.md** → know what to do
|
|
105
|
+
3. **Read boundary map** → know what's available from previous tasks
|
|
106
|
+
4. **Load ONLY files listed in PLAN.md** → don't read everything
|
|
107
|
+
5. **If you need a file not in PLAN.md** → note it, but question why
|
|
108
|
+
|
|
109
|
+
### Anti-Patterns
|
|
110
|
+
|
|
111
|
+
| ❌ Wrong | ✅ Right |
|
|
112
|
+
|----------|---------|
|
|
113
|
+
| Read entire codebase at start | Read STATE.md + task PLAN.md |
|
|
114
|
+
| Read all test files | Read only the test for current component |
|
|
115
|
+
| Read entire config files | Read only the section you need |
|
|
116
|
+
| Load the full design system | Load only the tokens file |
|
|
117
|
+
| Read all 22 agents at start | Read only the selected agent |
|
|
118
|
+
|
|
119
|
+
---
|
|
120
|
+
|
|
121
|
+
## 🧮 BUDGET ESTIMATION
|
|
122
|
+
|
|
123
|
+
Before starting a task, estimate:
|
|
124
|
+
|
|
125
|
+
```markdown
|
|
126
|
+
## Context Budget Estimate — T03
|
|
127
|
+
|
|
128
|
+
| Resource | Estimated | Budget |
|
|
129
|
+
|----------|-----------|--------|
|
|
130
|
+
| Files to read | 4 | < 8 ✅ |
|
|
131
|
+
| Files to write | 2 | < 8 ✅ |
|
|
132
|
+
| Lines to write | ~120 | < 400 ✅ |
|
|
133
|
+
| Steps | 5 | < 8 ✅ |
|
|
134
|
+
| New deps | 0 | < 3 ✅ |
|
|
135
|
+
|
|
136
|
+
**Verdict:** ✅ Fits in one context window.
|
|
137
|
+
```
|
|
138
|
+
|
|
139
|
+
If ANY metric is in 🔴 territory → SPLIT before starting.
|
|
140
|
+
|
|
141
|
+
---
|
|
142
|
+
|
|
143
|
+
## ⚡ QUICK REFERENCE
|
|
144
|
+
|
|
145
|
+
```
|
|
146
|
+
Task too big? → SPLIT
|
|
147
|
+
Too many files? → SPLIT
|
|
148
|
+
Too many steps? → SPLIT
|
|
149
|
+
Can't finish in one session? → SPLIT
|
|
150
|
+
|
|
151
|
+
SPLIT = create sub-tasks with:
|
|
152
|
+
- Own PLAN.md
|
|
153
|
+
- Own must-haves
|
|
154
|
+
- Boundary map connecting them
|
|
155
|
+
```
|
|
@@ -0,0 +1,278 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: context-gathering-patterns
|
|
3
|
+
description: >
|
|
4
|
+
Coleta de contexto com máxima eficiência de tokens. Usa pirâmide de custo
|
|
5
|
+
crescente: frontmatter → seções específicas → docs completos → Flyee API.
|
|
6
|
+
Obrigatório antes de qualquer implementação (SIMPLE e COMPLEX CODE).
|
|
7
|
+
skills:
|
|
8
|
+
- history-check-patterns
|
|
9
|
+
---
|
|
10
|
+
|
|
11
|
+
# Context Gathering Patterns
|
|
12
|
+
|
|
13
|
+
> **Regra de ouro:** Leia o mínimo necessário para tomar a decisão correta.
|
|
14
|
+
> Token = custo. Leitura desnecessária = contexto poluído + janela reduzida.
|
|
15
|
+
|
|
16
|
+
---
|
|
17
|
+
|
|
18
|
+
## 🎯 PROPÓSITO
|
|
19
|
+
|
|
20
|
+
Garantir que antes de implementar qualquer mudança o agente tenha contexto
|
|
21
|
+
suficiente — sem desperdiçar tokens com leituras redundantes.
|
|
22
|
+
|
|
23
|
+
**Dois objetivos em tensão:**
|
|
24
|
+
- ✅ Contexto completo (evitar bugs, não reinventar, seguir padrões)
|
|
25
|
+
- ✅ Economia de tokens (janela de contexto é recurso limitado)
|
|
26
|
+
|
|
27
|
+
**Solução:** Pirâmide de leitura com custo crescente. Parar no nível mais baixo
|
|
28
|
+
que responda às perguntas necessárias.
|
|
29
|
+
|
|
30
|
+
---
|
|
31
|
+
|
|
32
|
+
## 🔺 PIRÂMIDE DE CONTEXTO (custo crescente → parar o quanto antes)
|
|
33
|
+
|
|
34
|
+
```
|
|
35
|
+
Nível 1 — ~100 tokens → agent_context (YAML frontmatter dos docs relevantes)
|
|
36
|
+
Nível 2 — ~300 tokens → INDEX.md + tabelas Summary nos docs
|
|
37
|
+
Nível 3 — ~500 tokens → RETRO / LESSONS-LEARNED / BUGS-KNOWN
|
|
38
|
+
Nível 4 — ~800 tokens → Seção ESPECÍFICA do SDD/PRD (não o arquivo inteiro)
|
|
39
|
+
Nível 5 — ~1.500 tokens → Arquivo de código afetado + CODEBASE.md
|
|
40
|
+
Nível 6 — ~3.000 tokens → bridge --search-context (resultado filtrado, não lista)
|
|
41
|
+
Nível 7 — NUNCA → Histórico bruto de conversa / leitura de outros projetos
|
|
42
|
+
```
|
|
43
|
+
|
|
44
|
+
> [!CAUTION]
|
|
45
|
+
> **PROIBIDO:** Ler PRD ou SDD inteiros quando apenas uma seção é relevante.
|
|
46
|
+
> **PROIBIDO:** Usar histórico de conversa como fonte de contexto — use docs locais.
|
|
47
|
+
> **PROIBIDO:** Analisar código de outros projetos como referência — use padrões documentados.
|
|
48
|
+
|
|
49
|
+
---
|
|
50
|
+
|
|
51
|
+
## 🔍 QUANDO USAR CADA NÍVEL
|
|
52
|
+
|
|
53
|
+
### Request SIMPLE CODE (fix, add, change — arquivo único)
|
|
54
|
+
|
|
55
|
+
```
|
|
56
|
+
Nível 1 → Nível 2 → Nível 5
|
|
57
|
+
```
|
|
58
|
+
Ler: frontmatter do BREAKDOWN → Summary table → arquivo afetado.
|
|
59
|
+
Parar se já há resposta.
|
|
60
|
+
|
|
61
|
+
### Request COMPLEX CODE / DESIGN (build, create, implement)
|
|
62
|
+
|
|
63
|
+
```
|
|
64
|
+
Nível 1 → Nível 2 → Nível 3 → Nível 4 → Nível 5 → Nível 6 (se necessário)
|
|
65
|
+
```
|
|
66
|
+
Cada nível só é acionado se o anterior não respondeu as perguntas do Gate.
|
|
67
|
+
|
|
68
|
+
---
|
|
69
|
+
|
|
70
|
+
## 📋 PROCESSO PASSO A PASSO
|
|
71
|
+
|
|
72
|
+
### ► Passo 0: Verificar INDEX.md (sempre, custo mínimo)
|
|
73
|
+
|
|
74
|
+
```bash
|
|
75
|
+
# Ler apenas o INDEX.md — mapa de todos os docs existentes
|
|
76
|
+
view_file docs/INDEX.md
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
Objetivo: saber QUAIS docs existem antes de buscá-los.
|
|
80
|
+
Nunca buscar `find docs/ -name "*.md"` — use o INDEX.
|
|
81
|
+
|
|
82
|
+
---
|
|
83
|
+
|
|
84
|
+
### ► Passo 1: Ler frontmatter dos docs relevantes (Nível 1)
|
|
85
|
+
|
|
86
|
+
Para cada doc relevante identificado no INDEX.md, ler **apenas o frontmatter**
|
|
87
|
+
(linhas entre `---` no início do arquivo). O campo `agent_context` contém o
|
|
88
|
+
resumo semântico que o agente precisa sem ler o documento inteiro.
|
|
89
|
+
|
|
90
|
+
**Documentos prioritários (na ordem):**
|
|
91
|
+
|
|
92
|
+
| Doc | Onde | Responde |
|
|
93
|
+
|-----|------|----------|
|
|
94
|
+
| `BREAKDOWN-tasks.md` | `docs/` | O que já foi implementado? Qual padrão seguir? |
|
|
95
|
+
| `RETRO-{projeto}.md` | `docs/` | Quais bugs ocorreram? O que não fazer? |
|
|
96
|
+
| `SDD-{projeto}.md` | `docs/design/` | Qual arquitetura? Quais ADRs se aplicam? |
|
|
97
|
+
| `PRD-{projeto}.md` | `docs/` | Qual o escopo? Feature está aprovada? |
|
|
98
|
+
|
|
99
|
+
**Gate:** Se o `agent_context` do BREAKDOWN responde "o que existe" e "como fazer"
|
|
100
|
+
→ **pular Passo 2 e ir direto para Passo 4**.
|
|
101
|
+
|
|
102
|
+
> [!TIP]
|
|
103
|
+
> O `agent_context` bem escrito substitui 95% do History Check e 80% da leitura
|
|
104
|
+
> de SDD/PRD para tasks de implementação de features já mapeadas.
|
|
105
|
+
|
|
106
|
+
---
|
|
107
|
+
|
|
108
|
+
### ► Passo 2: Summary tables (Nível 2)
|
|
109
|
+
|
|
110
|
+
Se o frontmatter não foi suficiente, ler as tabelas Summary dos docs.
|
|
111
|
+
**Ler apenas as primeiras 80 linhas** dos docs de tracking (BREAKDOWN, PROGRESS).
|
|
112
|
+
|
|
113
|
+
```
|
|
114
|
+
view_file docs/BREAKDOWN-tasks.md (linhas 1-120, não o arquivo inteiro)
|
|
115
|
+
view_file docs/INDEX.md (completo — é sempre curto)
|
|
116
|
+
```
|
|
117
|
+
|
|
118
|
+
---
|
|
119
|
+
|
|
120
|
+
### ► Passo 3: RETRO / Lições (Nível 3)
|
|
121
|
+
|
|
122
|
+
Ler `docs/RETRO-{projeto}-v{N}.md` — documento de retrospectiva com bugs
|
|
123
|
+
conhecidos e decisões tomadas. É compacto por design.
|
|
124
|
+
|
|
125
|
+
**Padrão de busca dentro do RETRO:**
|
|
126
|
+
```bash
|
|
127
|
+
grep -n "<keyword da task>" docs/RETRO-*.md
|
|
128
|
+
```
|
|
129
|
+
|
|
130
|
+
Se o RETRO responder "já tentamos isso e o problema foi X" → não precisar ler SDD.
|
|
131
|
+
|
|
132
|
+
---
|
|
133
|
+
|
|
134
|
+
### ► Passo 4: Seção específica do SDD/PRD (Nível 4)
|
|
135
|
+
|
|
136
|
+
**NUNCA ler o SDD ou PRD inteiros.** Identificar a seção pelo índice e ler só ela.
|
|
137
|
+
|
|
138
|
+
```bash
|
|
139
|
+
# Primeiro: encontrar o número da linha da seção
|
|
140
|
+
grep -n "## <keyword>" docs/design/SDD-*.md | head -5
|
|
141
|
+
|
|
142
|
+
# Depois: ler apenas aquela seção (~50-100 linhas)
|
|
143
|
+
view_file docs/design/SDD-flyee.md --start <linha> --end <linha+80>
|
|
144
|
+
```
|
|
145
|
+
|
|
146
|
+
Seções mais acessadas por tipo de task:
|
|
147
|
+
|
|
148
|
+
| Tipo de task | Seção do SDD | Seção do PRD |
|
|
149
|
+
|---|---|---|
|
|
150
|
+
| Nova feature | Sec 2 (features) + Sec 13 (impl order) | Sec 6 (requirements) |
|
|
151
|
+
| SDK / endpoint | Sec 3.2 (SDK modules) | N/A |
|
|
152
|
+
| UI / componente | Sec 5 (component patterns) | N/A |
|
|
153
|
+
| Bug / comportamento | Sec 4 (data flow) + Sec 9 (error handling) | N/A |
|
|
154
|
+
| ADR / decisão | Sec 10 (ADRs) | N/A |
|
|
155
|
+
|
|
156
|
+
---
|
|
157
|
+
|
|
158
|
+
### ► Passo 5: Arquivo de código afetado (Nível 5)
|
|
159
|
+
|
|
160
|
+
Ler o arquivo específico que será alterado + verificar `CODEBASE.md` para
|
|
161
|
+
dependentes. **Não ler mais arquivos além do necessário.**
|
|
162
|
+
|
|
163
|
+
```bash
|
|
164
|
+
# Antes de ler código: verificar dependentes
|
|
165
|
+
grep -n "<nome-do-arquivo>" docs/CODEBASE.md
|
|
166
|
+
```
|
|
167
|
+
|
|
168
|
+
---
|
|
169
|
+
|
|
170
|
+
### ► Passo 6: Resource Discovery (Nível 6 — quando necessário)
|
|
171
|
+
|
|
172
|
+
Somente se Passos 0–5 não responderam às perguntas de implementação.
|
|
173
|
+
|
|
174
|
+
#### Caminho A — Resources Locais
|
|
175
|
+
|
|
176
|
+
**Condição:** `.agent/project-resources.json` existe.
|
|
177
|
+
|
|
178
|
+
```bash
|
|
179
|
+
cat .agent/project-resources.json # ler inteiro — é sempre pequeno
|
|
180
|
+
```
|
|
181
|
+
|
|
182
|
+
Para cada resource, comparar `scope[]` com keywords da task.
|
|
183
|
+
Match → `view_file` no `path` do resource.
|
|
184
|
+
|
|
185
|
+
#### Caminho B — Busca Semântica Flyee
|
|
186
|
+
|
|
187
|
+
**Condição:** `flyee.json` existe AND `enabled: true`.
|
|
188
|
+
|
|
189
|
+
```bash
|
|
190
|
+
python3 .agent/flyee-bridge/bridge.py --search-context "<keywords da task>"
|
|
191
|
+
```
|
|
192
|
+
|
|
193
|
+
Output: JSON com resultados semânticos. Ler apenas `results[]`, não a lista
|
|
194
|
+
completa de tasks.
|
|
195
|
+
|
|
196
|
+
> [!WARNING]
|
|
197
|
+
> **NÃO usar `--list-tasks`** para contexto — retorna apenas IDs sem conteúdo.
|
|
198
|
+
> Usar `--search-context` que retorna conteúdo semanticamente relevante.
|
|
199
|
+
|
|
200
|
+
---
|
|
201
|
+
|
|
202
|
+
## 🚨 ANTI-PATTERNS (com custo real de tokens)
|
|
203
|
+
|
|
204
|
+
| ❌ Anti-pattern | Custo | ✅ Correto |
|
|
205
|
+
|---|---|---|
|
|
206
|
+
| Ler PRD inteiro (~700 linhas) | ~8.000 tokens | Ler `agent_context` frontmatter (~80 tokens) |
|
|
207
|
+
| Ler SDD inteiro (~1.200 linhas) | ~15.000 tokens | Ler seção específica via grep |
|
|
208
|
+
| `bridge --list-tasks` para contexto | ~500 tokens de lista vazia | `agent_context` do BREAKDOWN |
|
|
209
|
+
| Histórico de conversa como contexto | Explode context window | RETRO + BREAKDOWN frontmatter |
|
|
210
|
+
| Analisar código de outros projetos | Ruído + tokens | `PADRÕES ESTABELECIDOS` no frontmatter |
|
|
211
|
+
| `find docs/ -name "*.md"` | Lento + irrelevante | Ler `docs/INDEX.md` |
|
|
212
|
+
| Ler arquivo de código sem verificar deps | Bug silencioso | CODEBASE.md primeiro |
|
|
213
|
+
|
|
214
|
+
> [!WARNING]
|
|
215
|
+
> **Caso real (Task #21):** Agente retomou conversa truncada e saltou Context
|
|
216
|
+
> Gathering. Inferiu tipos a partir do código, ignorando docs de fluxo.
|
|
217
|
+
> Resultado: decisão incorreta que causaria bug em runtime.
|
|
218
|
+
>
|
|
219
|
+
> **Prevenção:** Passo 0 (INDEX.md) + Passo 1 (frontmatter) tomam < 200 tokens
|
|
220
|
+
> e evitam este problema completamente.
|
|
221
|
+
|
|
222
|
+
---
|
|
223
|
+
|
|
224
|
+
## 📊 CHECKLIST DE EVIDÊNCIA (compacto)
|
|
225
|
+
|
|
226
|
+
```markdown
|
|
227
|
+
📖 CONTEXT — {título da task}
|
|
228
|
+
[ ] INDEX.md lido — docs existentes mapeados
|
|
229
|
+
[ ] Frontmatter BREAKDOWN lido — padrões e estado atual conhecidos
|
|
230
|
+
[ ] RETRO consultado (grep) — bugs conhecidos checados
|
|
231
|
+
[ ] Seção SDD relevante lida (ou N/A com justificativa)
|
|
232
|
+
[ ] Arquivo afetado + CODEBASE.md verificado
|
|
233
|
+
|
|
234
|
+
DECISÕES RELEVANTES: {bullet com as que impactam a task}
|
|
235
|
+
PADRÕES A SEGUIR: {bullet com SDK/hooks/styles confirmados}
|
|
236
|
+
RESTRIÇÕES: {bullet — o que NÃO fazer}
|
|
237
|
+
```
|
|
238
|
+
|
|
239
|
+
---
|
|
240
|
+
|
|
241
|
+
## ✅ GATE DE SAÍDA (mínimo para prosseguir)
|
|
242
|
+
|
|
243
|
+
Antes de escrever código, confirmar:
|
|
244
|
+
- [ ] Sei o que **já existe** no projeto (não vou reinventar)
|
|
245
|
+
- [ ] Sei qual **padrão** seguir (SDK, hooks, styles)
|
|
246
|
+
- [ ] Sei quais **restrições** se aplicam (ADRs, imports proibidos)
|
|
247
|
+
- [ ] Sei quais **arquivos** serão afetados e se há dependentes
|
|
248
|
+
|
|
249
|
+
Se algum desses pontos não foi respondido em Nível 1–4 → subir um nível.
|
|
250
|
+
Se nenhum doc responde → registrar: `"⚠️ Docs ausentes — implementar com base no padrão mais próximo e documentar decisão"`
|
|
251
|
+
|
|
252
|
+
---
|
|
253
|
+
|
|
254
|
+
## 🔄 RE-CHECK NO RESUME (conversa truncada)
|
|
255
|
+
|
|
256
|
+
Ao retomar uma conversa truncada:
|
|
257
|
+
|
|
258
|
+
1. Ir direto ao **Nível 1** (frontmatter do BREAKDOWN)
|
|
259
|
+
2. Se `agent_context` existe e está atualizado → **contexto restaurado em ~100 tokens**
|
|
260
|
+
3. Se RETRO existe → verificar bugs com `grep -n "<keyword>" docs/RETRO-*.md`
|
|
261
|
+
4. **NÃO tentar reler o histórico de conversa** — nunca é eficiente
|
|
262
|
+
|
|
263
|
+
> [!TIP]
|
|
264
|
+
> Um `agent_context` bem mantido no BREAKDOWN é a única ferramenta que garante
|
|
265
|
+
> retomada de contexto sem explodir a janela. É o "cache" do projeto.
|
|
266
|
+
|
|
267
|
+
---
|
|
268
|
+
|
|
269
|
+
## 🔗 WORKFLOWS QUE USAM ESTA SKILL
|
|
270
|
+
|
|
271
|
+
| Workflow | Fase | Nível mínimo |
|
|
272
|
+
|----------|------|-------------|
|
|
273
|
+
| `/execute` | Fase 4 (EXECUTE) | Nível 1–4 |
|
|
274
|
+
| `/new-task` | Fase 3 (EXECUTION) | Nível 1–4 |
|
|
275
|
+
| `/tdd` | Phase 6 (IMPLEMENT) | Nível 1–3 |
|
|
276
|
+
| `/new-project` | Phase 5 (CODE) | Nível 1–5 |
|
|
277
|
+
| `/legacy-project` | Phase 7B Passo 0 | Nível 1–5 |
|
|
278
|
+
| `/debug` | Qualquer fase | Nível 1–5 |
|