implement

Implements Specs and issues interactively with test-driven development and review-gated landing.

20|2|Updated May 16, 2026
One-click install
npx skills add https://github.com/reddb-io/red-skills --skill implement-reddb-io
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: implement
Source: https://github.com/reddb-io/red-skills/tree/main/plugins/dev/skills/engineering/implement
Command: npx skills add https://github.com/reddb-io/red-skills --skill implement-reddb-io

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Implementing a Spec or set of GitHub issues by hand often leads to skipped tests, work done on the wrong branch, and unreviewed merges. This Skill provides a guided, human-in-the-loop workflow that enforces test-first discipline, isolated worktrees, and review before landing. ## Core Features & Use Cases - Interactive Spec execution: Work through a Spec or issue set step by step with the agent guiding each slice, as the human-driven counterpart to the autonomous /afk fleet. - Test-first discipline: Invokes /tdd at pre-agreed seams, requiring a failing test before any implementation code, with typechecking after every cycle. - Safe worktree workflow: All work happens in a dedicated .red/tmp/worktrees/manual/<slug> worktree, never on the primary branch, and lands via /retake after /code-review passes. - Use Case: You have a Spec broken into GitHub issues and want to implement it yourself with agent guidance — the Skill walks you through scoping, TDD cycles, full test suite validation, code review, and landing the branch. ## Quick Start Ask the agent to implement the current Spec interactively, starting by agreeing the scope and breaking it into issues with /to-tickets.

Frequently Asked Questions about implement

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

FAQPage Schema
How do I implement a Spec interactively with an AI agent?▼

Use the /implement skill to work through a Spec or issue set step by step. It guides you through scoping, test-driven development cycles at agreed seams, typechecking, a final full test run, code review, and landing the branch via /retake.

What is the difference between /implement and /afk?▼

/implement is human-driven and interactive, covering one Spec or issue set in a manual worktree. /afk runs an autonomous fleet with no human in the loop, draining the full ready-for-agent queue with an isolated worktree per issue.

Can I implement directly on the main branch with this workflow?▼

No. All work must happen in a dedicated worktree under .red/tmp/worktrees/manual/<slug>, never on the primary branch or a sibling checkout. The branch is finished and landed through /retake after review passes.

When should I run the full test suite during implementation?▼

Run typechecking and isolated test files after every /tdd cycle, but run the full test suite only once at the end, after all issues in scope are individually green. Do not run /code-review while any test is red.

What happens if I skip writing a failing test first?▼

Skipping /tdd is treated as undefined behaviour for this skill. The workflow requires writing a failing test before any implementation code, then the minimum code to pass it, then refactoring.