implement

Implements approved specs and tickets using TDD with verification and code review gates.

15|3|Updated Jul 9, 2026
One-click install
npx skills add https://github.com/kiurakku/cursor-kit-for-ai --skill implement-kiurakku
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: implement
Source: https://github.com/kiurakku/cursor-kit-for-ai/tree/main/plugins/engineering/skills/implement
Command: npx skills add https://github.com/kiurakku/cursor-kit-for-ai --skill implement-kiurakku

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning an approved spec or ticket into shipped code often drifts in scope, skips tests, or ends without verification evidence. This Skill enforces a disciplined implementation loop so every acceptance criterion is coded, tested, and reviewed before being called done. ## Core Features & Use Cases - Contract-driven implementation: Loads spec acceptance criteria or ticket briefs as the source of truth and flags scope drift immediately. - TDD at agreed seams: Runs red-green-refactor per acceptance criterion with test seams confirmed before the first test is written. - Verification and review gates: Requires typecheck, targeted tests, and a code review pass on the diff, with a structured completion report recording exact commands and results. - Use Case: You have an agent-ready ticket to add a coupon API endpoint. The Skill walks you through confirming scope, writing failing tests per AC, implementing in dependency order, running pytest and typecheck, and producing a PR-ready completion report. ## Quick Start Ask the agent to implement ticket T-102 from the spec using TDD and produce a completion report with verification commands.

Frequently Asked Questions about implement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I implement a ticket with test-driven development?▼

Confirm the test seams with the user first, then run red-green-refactor per acceptance criterion: write a failing test, add minimal code to pass, refactor, and mark the AC done. Do not batch all tests at the end.

What should I check before starting implementation from a spec?▼

Verify you have an approved spec or agent-ready ticket, a branch from the latest main, no active S1 incident on the path, and secrets available via environment variables rather than pasted into code.

How do I handle scope creep during implementation?▼

Implement only the listed acceptance criteria or ticket objective. When new requirements appear, pause and update the spec with user confirmation instead of sneaking changes into the current work.

What verification is required before calling a ticket done?▼

Run the narrowest typecheck and test commands that prove the change, record the exact commands in the completion report, and pass a code review on the diff. Small diffs still require review since they can leak secrets or break auth.

What should I do when blocked mid-ticket?▼

Route by blocker type: update the spec for ambiguity, research or prototype unknown API behavior, debug unrelated CI failures without merging broken main, and write a handoff if the session ends mid-ticket.