qa

Files durable GitHub issues from conversational bug reports during interactive QA sessions.

Updated Mar 30, 2026
One-click install
npx skills add https://github.com/TrayMachi/dotfiles --skill qa-traymachi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: qa
Source: https://github.com/TrayMachi/dotfiles/tree/main/agents/skills/qa
Command: npx skills add https://github.com/TrayMachi/dotfiles --skill qa-traymachi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reporting bugs well is hard: users describe problems informally, and translating those reports into clear, actionable GitHub issues takes effort. This Skill runs an interactive QA session where the user describes issues conversationally while the agent explores the codebase for context and files well-structured GitHub issues automatically. ## Core Features & Use Cases - Conversational Issue Intake: Users describe bugs in their own words with at most 2-3 short clarifying questions, avoiding over-interviewing. - Background Codebase Exploration: An Explore subagent learns the project's domain language and expected behavior so issues use correct terminology without citing fragile file paths or line numbers. - Scope Assessment and Breakdown: Decides whether a report is a single issue or should be split into multiple thin, independently fixable issues with honest blocking relationships. - Use Case: During a QA pass on a web app, a user says "the form validation is wrong and the redirect after submit is broken." The agent files two separate GitHub issues with reproduction steps, marks dependencies, and shares the URLs. ## Quick Start Start a QA session and help me report the bugs I've been running into as GitHub issues.

Frequently Asked Questions about qa

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

FAQPage Schema
How do I file GitHub issues from a conversational bug report?

Describe the problem in your own words during the QA session. The agent asks at most 2-3 clarifying questions about expected versus actual behavior and reproduction steps, then files the issue with gh issue create and shares the URL.

How to write bug reports that survive code refactors?

Write issues from the user's perspective describing behaviors, not code. Avoid file paths, line numbers, and function names, and use the project's domain language so the issue stays accurate after major refactors.

When should a bug report be split into multiple issues?

Split when the fix spans independent areas, when concerns are separable enough for parallel work, or when there are distinct failure modes. Keep a single issue when one behavior is wrong in one place with a shared root cause.

Does this QA skill require the GitHub CLI?

Yes, issues are created with gh issue create, so the GitHub CLI must be installed and authenticated against the target repository. The agent files issues directly without asking for review first.

Why do the filed issues avoid mentioning file paths and line numbers?

File paths and line numbers go stale as the codebase changes, making issues confusing or misleading. Describing behaviors in domain language keeps issues durable and understandable after refactors.