Lesson 06 | Act 6: Wrap-up

⏱ Est. reading time: 4 min Updated on 5/7/2026

Act 6: Wrap-up (Episodes 13-14)

From "it works" to "delivery"

Episode 13: How Do We Know It's Done? — Verification Methodology

Scene: The code is written, but how do we confirm it's actually finished?

  • Verification is not just running tests once
  • Goal backtracking: Reverse-check from Phase goals, have all "must be true" statements been verified?
  • What /gsd-verify does: Independent agent reads PLAN.md goals → checks code → generates VERIFICATION.md
  • Code review: /gsd-code-review multi-dimensional scan (security/performance/maintainability), severity levels (Critical/High/Medium/Low)
  • Fix workflow: Review finds issues → /gsd-code-review-fix automatically patches line-by-line
  • What passing all 154 project tests means (what is covered, what isn't)
  • Core concept: Being done ≠ being right; verification is an independent step

Output Files:

  • .planning/phases/*/VERIFICATION.md — 4 verification reports
  • .planning/phases/*/REVIEW.md — Review reports

Command: /gsd-verify, /gsd-code-review

Episode 14: Retrospective — From a Sentence to a Complete Product

Scene: Wrapping up. Looking back at what we achieved along the way.

  • Start point: "Build a calculator that elderly people can use" (One sentence)
  • End point: 3 files / 1529 lines of code / 1631 lines of tests / 154 tests all passed
  • Complete file list and their roles:
    • index.html — 57 lines, semantic structure, full ARIA
    • style.css — 535 lines, CSS variable dual themes, 3-breakpoint responsiveness
    • app.js — 937 lines, four-layer architecture (engine/state machine/UI/persistence)
    • .planning/ — Full project planning logs (requirements/roadmap/decisions/plans/verification)
  • Bug hunt checklist: Floating-point precision, CSS Grid hidden elements, file:// protocol limitations
  • GSD Workflow recap:
    • /gsd-new-project — Idea → Structured
    • /gsd-discuss-phase — Lock decisions
    • /gsd-plan-phase — Break down tasks
    • /gsd-execute-phase — Write code
    • /gsd-verify — Verify
    • /gsd-quick — Quick fixes
  • Can this method be reused on other projects? (Yes — GSD applies to any software project)
  • v2 Outlook: Scientific calculation, currency conversion, PWA offline installation

Appendix

A. GSD Command Quick Reference Used in this Project

Command When to Use Times Used in Project
/gsd-new-project Have an idea, want to start 1
/gsd-discuss-phase Lock decisions before coding 4 (Once per Phase)
/gsd-plan-phase Break decisions into tasks 4
/gsd-execute-phase Execute tasks & code 4
/gsd-verify Confirm it was done right 4
/gsd-code-review Check code quality As needed
/gsd-quick Small fixes As needed
/gsd-debug Investigate bugs As needed

B. Technical Concept Index

Alphabetically sorted, noting which episode they appeared in:

  • ARIA labels → Episode 9
  • CSS Custom Properties → Episode 8
  • Event Delegation → Episode 6
  • FIFO Queue → Episode 5
  • Media Query → Episode 12
  • Shunting-yard → Episode 3
  • State Machine → Episode 7
  • TDD → Episode 4
  • XSS Prevention → Episode 11

Related Tools & Resources

Skill Marketplaces

A

Agent Skills

A platform providing a specification and showcase for AI agent skills, helping developers integrate and discover new capabilities for AI agents.
G

Google Agent Skills

A collection of agent skills for Google products and technologies, including Google Cloud, enabling AI agent developers to easily integrate and build intelligent applications.
A

Agent Skills

Production-grade engineering skills for AI coding agents, encoding workflows, quality gates, and best practices for software development.

Recommended Plugins

🏗️

Feature Dev

Comprehensive feature development workflow with a structured 7-phase approach. Agents: code-explorer, code-architect, code-reviewer.

Related Products

📦

LangGraph

LangGraph, from LangChain, Inc., is a powerful low-level orchestration framework engineered for building, managing, and deploying sophisticated, long-running, and stateful multi-agent applications. It represents application logic as cyclic Directed Acyclic Graphs (DAGs), giving developers precise control over LLM execution, including when to pause, loop, or transfer control. LangGraph ensures durable execution, seamless human-in-the-loop capabilities, and comprehensive short-term and long-term memory management. Deep integration with LangSmith provides robust debugging and production-ready deployment, making it the ideal choice for developing complex and reliable AI agents.
📦

deer-flow

DeerFlow (Deep Exploration and Efficient Research Flow) is an open-source super agent harness developed by ByteDance, designed to orchestrate sub-agents, memory, and sandboxes to perform a wide range of complex tasks. It is powered by an extensible skills system, allowing for custom functionality integration. DeerFlow 2.0 is a ground-up rewrite, offering enhanced agent capabilities and flexible deployment, recommended for use with large language models such as Doubao-Seed-2.0-Code, DeepSeek v3.2, and Kimi 2.5. It also integrates BytePlus's InfoQuest intelligent search and crawling toolset.
📦

dify

Dify is an open-source, production-ready LLM application development platform engineered to significantly accelerate the entire development cycle from prototype to production. It offers an intuitive visual workflow canvas, integrating a powerful Retrieval Augmented Generation (RAG) pipeline for efficient document ingestion and retrieval. Dify also features advanced agent capabilities based on LLM Function Calling or ReAct, configurable with pre-built or custom tools. It provides comprehensive support for hundreds of mainstream proprietary and open-source LLMs, a robust Prompt IDE for prompt engineering and model performance comparison, and built-in observability features (e.g., Opik, Langfuse, Arize Phoenix). This empowers developers to rapidly build, deploy, and monitor complex AI applications and agents.