skills.rest
whinchman
whinchmanCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent

Automates feature implementation from a todo file using TDD, git worktrees, and pull requests.

Updated May 2, 2026
Search Tags:#tdd#pull request#code generation#task automation#git worktree#autonomous agent#github cli
One-click install
npx skills add https://github.com/whinchman/midi-man-mk3 --skill agent-whinchman
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent
Source: https://github.com/whinchman/midi-man-mk3/tree/main/.claude/skills/agent
Command: npx skills add https://github.com/whinchman/midi-man-mk3 --skill agent-whinchman

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manually picking up backlog tasks, writing tests, implementing code, and opening pull requests is repetitive and error-prone. This Skill turns that entire loop into an autonomous workflow driven by a project configuration file. ## Core Features & Use Cases - Autonomous Task Pickup: Reads the next unchecked item from a todo file (or backlog) defined in agent.yaml and plans its implementation. - TDD Implementation Loop: Writes tests first, implements code, runs the full test suite, and commits each step in an isolated git worktree. - Pull Request Lifecycle: Opens PRs via the GitHub CLI, triages review comments and CI failures into new todo entries, and merges and cleans up after approval. - Use Case: A maintainer keeps a .workflow/TODO.md of features; invoking /agent picks the top item, implements it with tests on a feature branch, and opens a reviewable PR without touching the default branch. ## Quick Start Run /agent to pick up the next task from the todo file, implement it with tests, and open a pull request.

Frequently Asked Questions about agent

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

FAQPage Schema
How do I automate feature implementation from a todo list with an AI agent?▼

Define tasks in a todo markdown file and project conventions in agent.yaml, then invoke the /agent command. The agent picks the next unchecked item, plans it, implements it with tests in a git worktree, and opens a pull request.

How does the agent handle pull request review comments?▼

The triage step reads PR comments, reviews, and CI check results via the GitHub CLI. Each change request becomes a new todo entry with the comment text and URL as context, which the normal implementation pipeline then processes.

Can the agent work without writing tests?▼

Yes. Set testing.enabled to false in agent.yaml and the agent skips test writing and test runs, focusing on implementation and verifying correctness by reading the code for obvious errors.

Does the agent ever push directly to the main branch?▼

No. All work happens in a git worktree on a prefixed feature branch, and changes reach the default branch only through a pull request merged after approval. Direct pushes or merges to the default branch are explicitly forbidden.

What happens if the agent runs out of context mid-feature?▼

It completes the current plan step, commits the work, and records the stopping point in the plan file. A subsequent agent session resumes from that recorded step.

Related Skills

Explore More Software Engineering Skills→
product-on-purposeproduct-on-purpose
527

init-project-jpkb

Initialize JPKB projects with standardized documentation and folder structure.

Official
Basic
chainloop-devchainloop-dev
583

dependabot-pr-automation

Reviews, risk-assesses, approves, and merges open Dependabot pull requests on GitHub.

Official
Intermediate
erfanzarerfanzar
27

requesting-code-review

Automate pre-commit code review with static checks, linting, and auto-fixes.

Community
Advanced
2lab-ai2lab-ai

setup-ai-workflow

Configure an AI work-execution system with gate checks, commit/PR conventions, worktree isolation, and deployment runbook.

Official
Advanced