impl

Automate GitHub AFK issue implementation with /tdd cycles and commit actions.

16|9|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/sonthanh/brain-os-plugin --skill impl-sonthanh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: impl
Source: https://github.com/sonthanh/brain-os-plugin/tree/main/skills/impl
Command: npx skills add https://github.com/sonthanh/brain-os-plugin --skill impl-sonthanh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Automates the per-issue implementation workflow for brain-os tasks by taking AFK issues from GitHub and running the end-to-end /tdd cycle plus commit, push, and close actions, enabling autonomous progress on queued work.

Core Features & Use Cases

  • Once mode: pick a single AFK issue and run the /tdd cycle to implement it, then commit and close.
  • Issue mode: operate on a specific issue number regardless of owner label, bypassing the single-owner constraint.
  • Parallel mode: dispatch multiple AFK workers in parallel, each handling a distinct issue with per-issue cwd isolation.
  • Story mode: drain a multi-issue story in DAG order, coordinating AFK workers and HITL interactions until completion.
  • Auto mode: wrap /impl with a ralph-loop to drain the AFK backlog end-to-end, automating continuous processing until the backlog is exhausted or a limit is reached.
  • Per-issue worktree isolation and area-aware routing ensure commits land in the correct repository and reflect the intended area:* label mappings.

Quick Start

Type /impl to begin processing the next AFK issue, or /impl <N> to process a specific one.

Frequently Asked Questions about impl

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

FAQPage Schema
How do I automate GitHub issue processing with TDD cycles?

You can automate GitHub issue processing by pulling AFK issues and running an end-to-end TDD cycle that commits, pushes, and closes the work. This workflow supports autonomous progress on queued tasks without manual intervention.

How do I run parallel implementation workers on multiple GitHub issues?

Parallel mode dispatches multiple AFK workers simultaneously, each handling a distinct GitHub issue with per-issue cwd isolation. This allows separate worktrees and area-aware routing to process multiple issues concurrently without conflicts.

What is the best way to drain a multi-issue GitHub story in dependency order?

Story mode drains a multi-issue story in DAG order, coordinating AFK workers and human-in-the-loop interactions until completion. It ensures issues within a story are implemented sequentially according to their dependency graph.

Can I process a specific GitHub issue number without the single-owner constraint?

Yes, issue mode operates on a specific issue number regardless of owner label, bypassing the single-owner constraint. You simply invoke the command with the target issue number to process it directly.

How does per-issue worktree isolation prevent commit routing errors?

Per-issue worktree isolation enforces separate working directories and area-aware routing, ensuring commits land in the correct repository and reflect the intended area label mappings for each specific issue.

What are the limitations of automating continuous GitHub backlog processing?

Auto mode wraps the implementation process in a continuous loop to drain the AFK backlog, but processing stops when the backlog is exhausted or a trunk-block gate check fails, requiring explicit IMPL_AFK commits to proceed.