review-implement — for Claude Code review-implement, vibecode, community, for Claude Code, ide skills, ## Usage, lp:review-implement <plan_file_path>, Do NOT Modify, senior_pm, senior_uiux_designer

v1.0.0

Acerca de este Skill

Escenario recomendado: Ideal for AI agents that need full task guide : .agents/skills/tasks/review-implement/task.md. Resumen localizado: Fast mode (re-review trong loop) 1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong cùng 1 pass. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Características

Full task guide : .agents/skills/tasks/review-implement/TASK.md
/lp:review-implement <plan file path
Boundary Check — Verify no Do NOT Modify files were changed
Implementation Coverage Check — Verify delivered code still matches spec baseline + plan ACs
Multi-Persona Code Review — Review từ 4 góc nhìn bắt buộc trong 1 pass

# Temas principales

daudaudinang daudaudinang
[0]
[0]
Actualizado: 4/28/2026

Skill Overview

Start with fit, limitations, and setup before diving into the repository.

Escenario recomendado: Ideal for AI agents that need full task guide : .agents/skills/tasks/review-implement/task.md. Resumen localizado: Fast mode (re-review trong loop) 1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong cùng 1 pass. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

¿Por qué usar esta habilidad?

Recomendacion: review-implement helps agents full task guide : .agents/skills/tasks/review-implement/task.md. Fast mode (re-review trong loop) 1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong

Mejor para

Escenario recomendado: Ideal for AI agents that need full task guide : .agents/skills/tasks/review-implement/task.md.

Casos de uso accionables for review-implement

Caso de uso: Applying Full task guide : .agents/skills/tasks/review-implement/TASK.md
Caso de uso: Applying /lp:review-implement <plan file path
Caso de uso: Applying Boundary Check — Verify no Do NOT Modify files were changed

! Seguridad y limitaciones

  • Limitacion: Boundary Check — Verify no Do NOT Modify files were changed
  • Limitacion: Đã có ít nhất 1 lần review (NEEDS REVISION hoặc FAIL) trước đó → fast mode
  • Limitacion: Machine status chỉ dùng: PASS NEEDS REVISION FAIL

About The Source

The section below comes from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.

Demo Labs

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

FAQ y pasos de instalación

These questions and steps mirror the structured data on this page for better search understanding.

? Preguntas frecuentes

¿Qué es review-implement?

Escenario recomendado: Ideal for AI agents that need full task guide : .agents/skills/tasks/review-implement/task.md. Resumen localizado: Fast mode (re-review trong loop) 1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong cùng 1 pass. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

¿Cómo instalo review-implement?

Ejecuta el comando: npx killer-skills add daudaudinang/vibecode/review-implement. Funciona con Cursor, Windsurf, VS Code, Claude Code y más de 19 IDE adicionales.

¿Cuáles son los casos de uso de review-implement?

Los casos de uso principales incluyen: Caso de uso: Applying Full task guide : .agents/skills/tasks/review-implement/TASK.md, Caso de uso: Applying /lp:review-implement <plan file path, Caso de uso: Applying Boundary Check — Verify no Do NOT Modify files were changed.

¿Qué IDE son compatibles con review-implement?

Esta skill es compatible con Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer. Usa la CLI de Killer-Skills para una instalación unificada.

¿Tiene limitaciones review-implement?

Limitacion: Boundary Check — Verify no Do NOT Modify files were changed. Limitacion: Đã có ít nhất 1 lần review (NEEDS REVISION hoặc FAIL) trước đó → fast mode. Limitacion: Machine status chỉ dùng: PASS NEEDS REVISION FAIL.

Cómo instalar este skill

  1. 1. Abre tu terminal

    Abre la terminal o línea de comandos en el directorio de tu proyecto.

  2. 2. Ejecuta el comando de instalación

    Ejecuta: npx killer-skills add daudaudinang/vibecode/review-implement. La CLI detectará tu IDE o agente automáticamente y configurará la skill.

  3. 3. Empieza a usar el skill

    El skill ya está activo. Tu agente de IA puede usar review-implement de inmediato en el proyecto actual.

! Source Notes

This page is still useful for installation and source reference. Before using it, compare the fit, limitations, and upstream repository notes above.

Upstream Repository Material

The section below comes from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.

Upstream Source

review-implement

Fast mode (re-review trong loop) 1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong cùng 1 pass. This AI agent skill supports Claude

SKILL.md
Readonly
Upstream Repository Material
The section below comes from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.
Upstream Source

Review Implement

Full task guide: .agents/skills/tasks/review-implement/TASK.md

Usage

/lp:review-implement <plan_file_path>

What It Does

  1. Boundary Check — Verify no Do NOT Modify files were changed
  2. Implementation Coverage Check — Verify delivered code still matches spec baseline + plan ACs
  3. Multi-Persona Code Review — Review từ 4 góc nhìn bắt buộc trong 1 pass
  4. Finding Validation — Chỉ findings có evidence + context validation mới được dùng cho verdict
  5. Scoring & Findings — Chấm điểm định lượng và xuất machine status chuẩn

Execution model: Dual-mode review

Orchestrator chọn review mode dựa trên review history trong workflow:

Standard mode (lần review đầu tiên)

Spawn 4 agents độc lập, mỗi agent giữ đúng 1 persona, chạy song song:

PersonaPersona IDFocus
Senior PMsenior_pmFeature completeness, business scope adherence, missing requirements
Senior UI/UX Designersenior_uiux_designerUX fidelity so với plan, clarity, consistency, regression risk
Senior Developersenior_developerLogic correctness, tests/evidence, edge cases, touched-files/scope violations
System Architecturesystem_architectureCoupling, architecture impact, maintainability, technical-debt risk

Orchestrator validate evidence, normalize conflicts, tổng hợp verdict cuối.

Fast mode (re-review trong loop)

1 agent duy nhất chạy multi-persona — đánh giá lần lượt từ 4 góc nhìn trong cùng 1 pass.

  • Tập trung vào delta changes so với review trước
  • Đọc previous review output từ .codex/pipeline/PLAN_<NAME>/ để biết findings cũ
  • Nhanh hơn vì chỉ spawn 1 agent thay vì 4

Khi nào dùng mode nào?

  • Chưa có lần review nào trước đó trong workflow → standard mode
  • Đã có ít nhất 1 lần review (NEEDS_REVISION hoặc FAIL) trước đó → fast mode
  • Không được bỏ qua persona nào ở cả 2 modes

Quantitative verdict model

  • Mỗi persona chấm đúng 4 criteria cốt lõi, mỗi criteria 0..10
  • Điểm persona = trung bình 4 criteria
  • Điểm tổng = weighted average:
    • senior_developer: 30%
    • system_architecture: 30%
    • senior_pm: 20%
    • senior_uiux_designer: 20%
  • Machine status chỉ dùng: PASS | NEEDS_REVISION | FAIL

Fail-fast gates

Phải FAIL hoặc reject contract ngay nếu có một trong các lỗi sau:

  • thiếu bất kỳ mandatory persona nào trong output
  • finding không có evidence tối thiểu
  • thiếu business-context validation
  • Do NOT Modify scope violation
  • findings conflict nhưng chưa normalize/xác thực xong mà vẫn chốt PASS

Output

  • Review report: .codex/pipeline/<PLAN_NAME>/04-review-implement.output.md
  • Output contract: .codex/pipeline/<PLAN_NAME>/04-review-implement.output.contract.json

Orchestrator Reference

When Orchestrator receives /lp:review-implement, orchestrator phải coi đây là worker review step canonical.

Mandatory roster:

  • Task file: .agents/skills/tasks/review-implement/TASK.md
  • Personas bắt buộc: senior_pm, senior_uiux_designer, senior_developer, system_architecture

Decision authority:

  • Orchestrator vẫn giữ quyền quyết định cách spawn chi tiết.
  • Nhưng canonical flow không được bỏ qua persona nào trong 4 persona bắt buộc.
  • Nếu LP top-level flow gọi review step này, worker phải chạy qua agent trong current workspace theo policy của lp-pipeline-orchestrator, không mặc định isolate bằng worktree.

Notes

  • Skill này mô tả canonical review model cho review-implement, không phải roster gợi ý mềm.
  • Source of truth cho top-level LP spawn policy vẫn là .agents/skills/lp-pipeline-orchestrator/SKILL.md.
  • Human report có thể dùng wording dễ đọc, nhưng machine source of truth chỉ là PASS | NEEDS_REVISION | FAIL.
  • Khi review Epic phase, ưu tiên context từ direct dependency reports và phase report hiện tại; không dùng raw notes ngoài dependency trực tiếp.
  • Nếu plan bật broad_context_reports = true và tổng phase <= 5, cho phép mở rộng review context bằng completed reports theo thứ tự phase index tăng dần.
  • Nếu phát hiện branch dependency_critical = true thiếu dependencies explicit, report blocker waiting_user, không đề xuất fallback tuyến tính.

Habilidades relacionadas

Looking for an alternative to review-implement or another community skill for your workflow? Explore these related open-source skills.

Ver todo

openclaw-release-maintainer

Logo of openclaw
openclaw

Resumen localizado: 🦞 # OpenClaw Release Maintainer Use this skill for release and publish-time workflow. It covers ai, assistant, crustacean workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

333.8k
0
Inteligencia Artificial

widget-generator

Logo of f
f

Resumen localizado: Generate customizable widget plugins for the prompts.chat feed system # Widget Generator Skill This skill guides creation of widget plugins for prompts.chat . It covers ai, artificial-intelligence, awesome-list workflows. This AI agent skill supports Claude Code, Cursor, and

149.6k
0
Inteligencia Artificial

flags

Logo of vercel
vercel

Resumen localizado: The React Framework # Feature Flags Use this skill when adding or changing framework feature flags in Next.js internals. It covers blog, browser, compiler workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

138.4k
0
Navegador

pr-review

Logo of pytorch
pytorch

Resumen localizado: Usage Modes No Argument If the user invokes /pr-review with no arguments, do not perform a review . It covers autograd, deep-learning, gpu workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

98.6k
0
Desarrollador