name: bug-reproducer
description: "Create minimal, reliable reproduction of reported bugs from issues, logs, and user reports. Produces step-by-step reproduction cases with expected vs actual behavior."
phase: INIT
category: core
version: "1.0.0"
depends_on: []
tags: [debugging, reproduction, bugs, triage]

Bug Reproducer

Create minimal, reliable reproduction cases from bug reports, failing tests, user-reported issues, and production error logs. The goal is to distill a noisy bug report into a precise, repeatable set of steps that any engineer can follow to observe the defect.

When to Use

• A user or QA submits a bug report with incomplete reproduction steps
• A test is failing intermittently and the cause is unclear
• Production logs show errors that need local reproduction
• A reported issue needs verification before prioritization
• You need to confirm a bug still exists before investing in a fix

Process

1. Read the bug report or issue - Gather all available information: user description, screenshots, error messages, stack traces, logs, and environment details.
2. Identify symptoms and environment - Catalog the observable symptoms (error messages, incorrect behavior, crashes) and the environment where they occur (OS, browser, runtime version, configuration).
3. Create minimal reproduction steps - Strip away everything unnecessary. Find the shortest path from a clean state to observing the bug. Each step should be a concrete, unambiguous action.
4. Verify reproduction is reliable - Run the reproduction steps at least 3 times. If the bug is intermittent, document the success rate (e.g., "reproduces 2 out of 5 attempts") and any timing or ordering sensitivities.
5. Document expected vs actual behavior - State clearly what should happen at each critical step and what actually happens instead. Include exact error messages, return values, or visual artifacts.
6. Identify related code paths - Trace the reproduction steps to the relevant source files, functions, and lines. This primes the next phase (root cause analysis) with starting points.

Deliverables

BUG-REPRODUCTION.md Contents

• Summary: One-sentence description of the defect
• Environment: OS, runtime, versions, configuration, dependencies
• Steps to Reproduce: Numbered, concrete steps from clean state to bug
• Expected Behavior: What should happen
• Actual Behavior: What actually happens (with error output)
• Error Output: Full stack traces, logs, or screenshots
• Related Code Paths: Source files, functions, and lines involved
• Frequency/Reliability: How often the bug reproduces (e.g., 3/3, 2/5)
• Workarounds: Any known workarounds, if applicable

Quality Criteria

• Reproduction is reliable: reproduces at least 3 out of 3 attempts (or intermittent rate is documented)
• Reproduction is minimal: uses the fewest steps possible to trigger the bug
• Environment details are specific: exact versions, not ranges or "latest"
• Steps are unambiguous: another engineer can follow them without clarification
• Expected vs actual behavior is clearly separated
• Error output is complete (not truncated)
• Related code paths are identified (at least one file/function)