coding workflow

Bug Report Cleaner

Produce a scoped bug report with evidence, reproduction steps, expected behavior, and an acceptance test.

Setup: 10 minutes
Difficulty: Beginner
Tested: 2026-07-26

# Bug Report Cleaner

Turn a vague issue into a reproducible, developer-ready report.

**Outcome:** Produce a scoped bug report with evidence, reproduction steps, expected behavior, and an acceptance test.

**Category:** coding

**Setup:** 10 minutes · Beginner

## Requirements

- Original bug report or support message
- Relevant logs, screenshots, and environment details

**Tools:** cursor, github-copilot

## Expected output

A reproducible issue brief and regression-test outline

## Steps

### 1. Separate facts from guesses

Extract only observed behavior, environment facts, timestamps, and attached evidence.

**Prompt**

```text
Rewrite the report into Observed, Environment, Evidence, and Unknowns. Do not infer a root cause.
```

### 2. Build minimal reproduction

Propose the shortest reproducible sequence and mark every missing prerequisite as a question.

**Prompt**

```text
Create numbered reproduction steps using only known facts. List blocking questions separately.
```

### 3. Define expected behavior

State the expected result and the smallest pass/fail acceptance condition.

**Prompt**

```text
Write Expected Behavior and one regression-test scenario in Given/When/Then form.
```

### 4. Prepare the handoff

Combine the verified sections and attach links to evidence without adding diagnosis.

**Prompt**

```text
Produce a concise engineering handoff with severity left unassigned unless evidence supports it.
```

## Limitations

This playbook improves issue quality but does not diagnose or fix the underlying defect.

## Privacy

Remove credentials, access tokens, customer identifiers, and private production data before using an AI tool.

## Sources

- https://www.cursor.com/
- https://github.com/features/copilot

Tested 2026-07-26 · Updated 2026-07-26

# Bug Report Cleaner

Turn a vague issue into a reproducible, developer-ready report.

**Outcome:** Produce a scoped bug report with evidence, reproduction steps, expected behavior, and an acceptance test.

**Category:** coding

**Setup:** 10 minutes · Beginner

## Requirements

- Original bug report or support message
- Relevant logs, screenshots, and environment details

**Tools:** cursor, github-copilot

## Expected output

A reproducible issue brief and regression-test outline

## Steps

### 1. Separate facts from guesses

Extract only observed behavior, environment facts, timestamps, and attached evidence.

**Prompt**

```text
Rewrite the report into Observed, Environment, Evidence, and Unknowns. Do not infer a root cause.
```

### 2. Build minimal reproduction

Propose the shortest reproducible sequence and mark every missing prerequisite as a question.

**Prompt**

```text
Create numbered reproduction steps using only known facts. List blocking questions separately.
```

### 3. Define expected behavior

State the expected result and the smallest pass/fail acceptance condition.

**Prompt**

```text
Write Expected Behavior and one regression-test scenario in Given/When/Then form.
```

### 4. Prepare the handoff

Combine the verified sections and attach links to evidence without adding diagnosis.

**Prompt**

```text
Produce a concise engineering handoff with severity left unassigned unless evidence supports it.
```

## Limitations

This playbook improves issue quality but does not diagnose or fix the underlying defect.

## Privacy

Remove credentials, access tokens, customer identifiers, and private production data before using an AI tool.

## Sources

- https://www.cursor.com/
- https://github.com/features/copilot

Tested 2026-07-26 · Updated 2026-07-26

{
  "slug": "bug-report-cleaner",
  "title": "Bug Report Cleaner",
  "description": "Turn a vague issue into a reproducible, developer-ready report.",
  "outcome": "Produce a scoped bug report with evidence, reproduction steps, expected behavior, and an acceptance test.",
  "category": "coding",
  "tags": [
    "bugs",
    "triage",
    "quality"
  ],
  "setupMinutes": 10,
  "difficulty": "Beginner",
  "featured": true,
  "testedAt": "2026-07-26",
  "updatedAt": "2026-07-26",
  "tools": [
    "cursor",
    "github-copilot"
  ],
  "inputs": [
    "Original bug report or support message",
    "Relevant logs, screenshots, and environment details"
  ],
  "expectedOutput": "A reproducible issue brief and regression-test outline",
  "steps": [
    {
      "title": "Separate facts from guesses",
      "instruction": "Extract only observed behavior, environment facts, timestamps, and attached evidence.",
      "prompt": "Rewrite the report into Observed, Environment, Evidence, and Unknowns. Do not infer a root cause."
    },
    {
      "title": "Build minimal reproduction",
      "instruction": "Propose the shortest reproducible sequence and mark every missing prerequisite as a question.",
      "prompt": "Create numbered reproduction steps using only known facts. List blocking questions separately."
    },
    {
      "title": "Define expected behavior",
      "instruction": "State the expected result and the smallest pass/fail acceptance condition.",
      "prompt": "Write Expected Behavior and one regression-test scenario in Given/When/Then form."
    },
    {
      "title": "Prepare the handoff",
      "instruction": "Combine the verified sections and attach links to evidence without adding diagnosis.",
      "prompt": "Produce a concise engineering handoff with severity left unassigned unless evidence supports it."
    }
  ],
  "limitations": "This playbook improves issue quality but does not diagnose or fix the underlying defect.",
  "privacy": "Remove credentials, access tokens, customer identifiers, and private production data before using an AI tool.",
  "sources": [
    "https://www.cursor.com/",
    "https://github.com/features/copilot"
  ],
  "related": [
    "pr-review-pack"
  ]
}

Requirements

Bring these inputs.

Original bug report or support message
Relevant logs, screenshots, and environment details

Output contract

Know what done looks like.

A reproducible issue brief and regression-test outline

Execution

Run the playbook.

01
Separate facts from guesses
Extract only observed behavior, environment facts, timestamps, and attached evidence.
```
Rewrite the report into Observed, Environment, Evidence, and Unknowns. Do not infer a root cause.
```
02
Build minimal reproduction
Propose the shortest reproducible sequence and mark every missing prerequisite as a question.
```
Create numbered reproduction steps using only known facts. List blocking questions separately.
```
03
Define expected behavior
State the expected result and the smallest pass/fail acceptance condition.
```
Write Expected Behavior and one regression-test scenario in Given/When/Then form.
```
04
Prepare the handoff
Combine the verified sections and attach links to evidence without adding diagnosis.
```
Produce a concise engineering handoff with severity left unassigned unless evidence supports it.
```

Field note

How to use this route.

Use this playbook before asking a coding agent to repair anything. A clean report reduces speculative edits and gives the eventual fix a testable finish line.

Bring these inputs.

Know what done looks like.

Run the playbook.

Separate facts from guesses

Build minimal reproduction

Define expected behavior

Prepare the handoff

How to use this route.

Keep moving through the atlas