/scout

Scout Output Format

Purpose: Canonical investigation report schema, toolkit, and completion criteria. Read when: You are producing the final investigation artifact or checking whether the handoff is complete enough.

Contents

• Investigation report template
• Investigation toolkit
• Completion criteria
• Confidence levels

Investigation Report Template

## Scout Investigation Report

### Bug Summary
**Title:** [Brief description]
**Severity:** Critical / High / Medium / Low
**Reproducibility:** Always / Sometimes / Rare

### Reproduction Steps
1. [Step 1]
2. [Step 2]

**Expected:** [What should happen]
**Actual:** [What actually happens]

### Root Cause Analysis
**Location:** `src/path/to/file.ts:123` in `functionName()`
**Cause:** [Why the bug occurs]

### Recommended Fix
**Approach:** [High-level fix strategy]
**Files to modify:** [List with expected changes]

### Regression Prevention
**Suggested tests for Radar:** [Test cases to prevent recurrence]

## LLM Fix Prompt
[Paste-ready instruction prompt for a downstream coding LLM. Mandatory for confirmed root causes. See `reference/fix-prompt-generation.md` for verbs, schema, and suppression rules.]

Add when available:

• confidence level
• evidence links
• impact scope
• workaround

Recommended Fix Impact Scope Template

RecommendedFixImpactScope:
  callers:    {affected: [file:line, ...], note: "1-line description or 'none'"}
  tests:      {affected: [test files], note: "additions/updates needed or 'none'"}
  types:      {affected: [type/schema files], note: "contract impact or 'none'"}
  configs:    {affected: [config/env keys], note: "propagation impact or 'none'"}
  docs:       {affected: [doc paths], note: "update needed or 'none'"}
  axes_affected: <integer 0-5>
  recommend_ripple: <true if axes_affected >= 3 OR uncertainty is high>

Investigation Toolkit

Investigation Completion Criteria

Required

• Reproducible, or reproduction conditions identified
• Root cause identified, or a bounded hypothesis set exists
• Impact scope understood
• Fix approach can be articulated

Confidence Levels