Expert guide for automated and manual Web Accessibility (a11y) testing — axe-core, Pa11y, Playwright a11y, screen reader testing, and WCAG 2.2 Level AA/AAA compliance / Panduan ahli pengujian aksesibilitas web.
日本語の概要は準備中です。原文の説明を表示しています。
Updated to cover both pre-code PRD/ERD/Docs generation AND post-code continuous documentation updates / Diperbarui untuk mencakup pembuatan PRD/ERD/Docs pra-kode DAN pembaruan dokumentasi kontinu pasca-kode.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
<a name="english"></a>
Connects with anti-slop, brainstorming, zero-to-prod-orchestrator, session-memory-manager to ensure cohesive execution and zero-slop documentation updates.
A mandatory lifecycle guardrail covering two distinct phases:
CHANGELOG.md, BLUEPRINT.md, PROGRESS.md, and Architecture Decision Records (ADRs).Why PRD Before Code: Prevents scope creep, aligns AI output, enables traceability, reduces rework, version-controlled.
PRD.md, ERD.md, DOKUMENTASI.md, and ROADMAP.md (or PROGRESS.md). Enforce anti-slop: zero marketing buzzwords; mandate concrete PostgreSQL DDL schemas, explicit REST/RPC contracts, and verified latency NFR budgets.# Product Requirements Document (PRD)
**Project**: [Project Name]
**Version**: 1.0.0
**Status**: Draft | In Review | Approved
**Created**: YYYY-MM-DD
**Last Updated**: YYYY-MM-DD
---
## 1. Executive Summary
[2-3 sentences: What is this product? Who is it for? What problem does it solve?]
## 2. Problem Statement
**Problem**: [Clear description]
**Target Users**: [Specific segments]
**Current Pain Points**: [List]
## 3. Goals & Success Metrics
| Goal | Metric | Target |
|---|---|---|
| [Goal] | [Metric] | [Target] |
## 4. User Personas
### Persona 1: [Name]
- **Role**: [Job title]
- **Goals**: [What they want to achieve]
- **Frustrations**: [What currently doesn't work]
- **Key Behaviors**: [How they'll use this product]
## 5. Feature Requirements
### MVP Features (Must Have — v1.0)
- [ ] **[Feature Name]**: [Description. Acceptance criteria: ...]
### Phase 2 Features (Should Have — v1.x)
- [ ] **[Feature Name]**: [Description]
### Future Features (Nice to Have — v2.0+)
- [ ] **[Feature Name]**: [Description]
## 6. Technical Architecture
### Stack Decision
| Layer | Technology | Rationale |
|---|---|---|
| [Layer] | [Technology] | [Reason] |
### Architecture Decisions (ADRs)
- **ADR-001**: [Decision title] — [Decision made and why]
### Multi-Entry Points (if SaaS)
| Entry Point | Domain | Purpose |
|---|---|---|
| [Entry] | [Domain] | [Purpose] |
## 7. Data Model (High-Level)
users ──belongs_to──> workspaces (via workspace_members)
workspaces ──has_many──> projects
## 8. User Flows
### Primary Flow: [Name]
1. User [action 1]
2. System [response 1]
## 9. Non-Functional Requirements
| Requirement | Target |
|---|---|
| Performance | LCP < 2.5s, INP < 200ms |
## 10. Out of Scope
- [Explicitly excluded item]
## 11. Open Questions
- [ ] [Question that needs decision]
## 12. Approval & Sign-off
| Stakeholder | Role | Status |
|---|---|---|
| [Name] | [Role] | [Status] |
After every successful feature implementation:
[Unreleased] (Keep-a-Changelog standard: Added/Changed/Fixed/Removed/Security).[x], update next steps.docs/adr/ADR-NNN-title.md) if a significant architectural decision was made and update index (docs/adr/README.md).| File | Purpose | Update Frequency |
|---|---|---|
CHANGELOG.md | User-facing list of changes | Every PR / feature |
BLUEPRINT.md | Technical architecture overview | Major structural changes |
PROGRESS.md | Development roadmap and status | Each work session |
docs/adr/ | Architecture Decision Records | Each key decision |
# ADR-001: [Title]
**Status**: Accepted | Rejected | Proposed | Deprecated
**Date**: YYYY-MM-DD
**Deciders**: [Team/Individual]
## Context
[What motivated this decision?]
## Decision
[What is being decided?]
## Rationale
[Why this choice?]
## Consequences
**Positive:** / **Negative:**
## Superseded By
[ADR-XXX] (if applicable)
MANDATORY: When building a new application, use generate_image to create a custom logo. Use it as:
favicon.ico).og:image).<a name="bahasa-indonesia"></a>
Terhubung dengan brainstorming, zero-to-prod-orchestrator, session-memory-manager.
Guardrail siklus hidup wajib:
CHANGELOG.md, BLUEPRINT.md, PROGRESS.md, dan ADR.PRD.md, ERD.md, DOKUMENTASI.md, ROADMAP.md → 4. Tinjau & setujui → 5. Baru mulai kode → 6. Referensi selalu.[Unreleased] (Added/Changed/Fixed/Removed/Security).[x], perbarui langkah selanjutnya.docs/adr/ADR-NNN-title.md untuk keputusan besar.WAJIB: Gunakan generate_image untuk membuat logo kustom sebagai logo, favicon, dan OG image.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Expert guide for automated and manual Web Accessibility (a11y) testing — axe-core, Pa11y, Playwright a11y, screen reader testing, and WCAG 2.2 Level AA/AAA compliance / Panduan ahli pengujian aksesibilitas web.
日本語の概要は準備中です。原文の説明を表示しています。
Expert guide for intelligent model cascading and routing — complexity-scored task routing from Flash/Haiku to Sonnet/Opus/Astra, dynamic escalation with quality gates, 40-60% token cost reduction while maintaining output quality / Panduan ahli untuk kaskade dan routing model cerdas — routing tugas berbasis skor kompleksitas dari Flash/Haiku ke Sonnet/Opus/Astra, eskalasi dinamis dengan gerbang kualitas, pengurangan biaya token 40-60% dengan kualitas output terjaga.
日本語の概要は準備中です。原文の説明を表示しています。
Expert guide for Affective Computing, emotional AI, and real-time sentiment analysis through native multimodal tokens (voice intonation and facial micro-expressions) / Panduan ahli komputasi afektif, AI emosional, dan analisis sentimen real-time melalui token multimodal native.
日本語の概要は準備中です。原文の説明を表示しています。
Expert guide for AI-assisted coding workflows — agentic code generation, multi-agent code swarms, self-healing CI/CD, automated PR review, spec-to-code pipelines, codebase knowledge graphs, and human-in-the-loop approval gates / Panduan ahli untuk workflow pengkodean berbasis AI — generasi kode agentic, code swarm multi-agen, CI/CD self-healing, review PR otomatis, pipeline spec-to-code, knowledge graph codebase, dan gate persetujuan human-in-the-loop.
日本語の概要は準備中です。原文の説明を表示しています。
Expert guide for long-term episodic memory integration (Mem0 v2, Letta/MemGPT, Zep v2), memory tier architecture, pgvector HNSW storage, and unified context management for autonomous AI agents / Panduan ahli untuk integrasi memori episodik jangka panjang (Mem0 v2, Letta/MemGPT, Zep v2), arsitektur tier memori, penyimpanan pgvector HNSW, dan manajemen konteks terpadu untuk agen AI otonom.
日本語の概要は準備中です。原文の説明を表示しています。
Expert guide for designing Machine-to-Machine (M2M) micro-economies, autonomous agent wallets, and swarm budget allocation / Panduan ahli merancang ekonomi mikro antar-agen (M2M), dompet agen otonom, dan alokasi anggaran swarm.
日本語の概要は準備中です。原文の説明を表示しています。