bug-reproduction-brief

Converts vague bug reports into minimal evidence-backed reproductions before any fix is attempted.

38.5k|4.9k|Updated Jun 11, 2025
One-click install
npx skills add https://github.com/github/awesome-copilot --skill bug-reproduction-brief
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bug-reproduction-brief
Source: https://github.com/github/awesome-copilot/tree/main/skills/bug-reproduction-brief
Command: npx skills add https://github.com/github/awesome-copilot --skill bug-reproduction-brief

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Bug reports are often vague, intermittent, environment-specific, or mixed with an assumed cause, leading developers to fix the wrong thing or destroy evidence while investigating. This Skill enforces a disciplined workflow that proves the smallest observable failure before any diagnosis or repair begins.

Core Features & Use Cases

  • Structured Evidence Capture: Records the exact error, environment facts (commit, runtime versions, lockfile, feature flags), and separates expected from actual behavior without embedding suspected causes.
  • Minimal Reproduction Reduction: Systematically removes unrelated data, services, and steps until only the smallest failing fixture remains, then proves repeatability by running it multiple times.
  • Safety Boundaries: Prohibits editing implementation code during reproduction, changing production data, or exposing secrets, keeping diagnosis separate from remediation.
  • Use Case: A QA engineer reports that checkout fails intermittently on staging. Use this Skill to reduce the failure to a minimal test fixture, document the exact commands and outputs, and hand a verified reproduction brief to the developer.

Quick Start

Use the Bug Reproduction Brief skill on the failing checkout test, reduce it to the smallest safe failing fixture, and report the exact command evidence, expected result, actual result, and remaining unknowns without fixing it.

Frequently Asked Questions about bug-reproduction-brief

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

FAQPage Schema
How do I reproduce a bug that only happens intermittently?

Run the minimal reproduction multiple times where safe and record the observed frequency and duration instead of calling it deterministic. Reduce the failing path step by step, restoring the last removed condition if the failure stops, and document exact commands and outputs.

How to write a minimal bug reproduction before fixing code?

Capture the exact error and smallest known input, record environment facts like commit and runtime versions, then state expected versus actual behavior separately. Remove unrelated data and steps one at a time until only the smallest failing fixture remains, and stop before editing any implementation code.

What environment details should a bug report include?

Include only inspectable facts: repository and commit, runtime and package-manager versions, operating system or container, dependency lockfile, relevant feature flags, and whether the target is local, test, staging, or production. Never guess credentials or production configuration.

When should I stop investigating a bug and start fixing it?

Stop once you have a verified, repeatable reproduction with documented evidence. Editing implementation code during reproduction can destroy evidence and mix diagnosis with remediation, so repair is treated as a separate workflow.

Can I reproduce a bug directly in production?

No. Prefer an isolated test, minimal script, or smallest safe request over reproducing against production, and never change production data merely to reproduce a bug. Use read-only or reversible discovery methods first.