# Evidence Requirements Task

Enforce evidence-based QA with mandatory proof of fix and verification.

**Absorbed from:** Auto-Claude PR Review Phase 3 - Evidence Requirements

---

## Task Definition

```yaml
task: qaEvidenceRequirements()
responsavel: Quinn (Guardian)
atomic_layer: Molecule

inputs:
  - story_id: string (required)
  - issue_type: "bug_fix" | "feature" | "dependency_update" | "refactor"

outputs:
  - evidence_checklist: file (docs/stories/{story-id}/qa/evidence_checklist.md)
  - evidence_status: object { complete: boolean, missing: string[] }
```

---

## Evidence Checklists by Issue Type

### Bug Fix Evidence

```yaml
bug_fix_evidence:
  required:
    - id: original-error
      name: 'Original Error Documented'
      description: 'Screenshot, log, or reproduction steps of the bug'
      severity: CRITICAL

    - id: root-cause
      name: 'Root Cause Identified'
      description: 'Clear explanation of why the bug occurred'
      severity: HIGH

    - id: before-after
      name: 'Before/After Comparison'
      description: 'Code diff showing the fix with explanation'
      severity: HIGH

    - id: regression-test
      name: 'Regression Test Added'
      description: 'Test case that would catch this bug if reintroduced'
      severity: CRITICAL

  optional:
    - id: related-issues
      name: 'Related Issues Checked'
      description: 'Similar code patterns checked for same bug'
      severity: LOW
```

### Feature Implementation Evidence

```yaml
feature_evidence:
  required:
    - id: acceptance-verified
      name: 'All Acceptance Criteria Verified'
      description: 'Each criterion has proof of completion'
      severity: CRITICAL

    - id: edge-cases
      name: 'Edge Cases Tested'
      description: 'Boundary conditions and error states verified'
      severity: HIGH

    - id: happy-path
      name: 'Happy Path Demonstrated'
      description: 'Primary use case works as expected'
      severity: CRITICAL

  conditional:
    - id: cross-platform
      name: 'Cross-Platform Tested'
      condition: 'feature has UI component'
      platforms: ['Chrome', 'Firefox', 'Safari', 'Mobile']
      severity: MEDIUM

    - id: performance-impact
      name: 'Performance Impact Assessed'
      condition: 'feature is performance-critical'
      severity: HIGH
```

### Dependency Update Evidence

```yaml
dependency_evidence:
  required:
    - id: security-check
      name: 'Security Vulnerabilities Checked'
      description: 'npm audit or equivalent shows no new vulnerabilities'
      severity: CRITICAL

    - id: license-check
      name: 'License Compatibility Verified'
      description: 'New dependency license is compatible with project'
      severity: HIGH

    - id: breaking-changes
      name: 'Breaking Changes Handled'
      description: 'Changelog reviewed and breaking changes addressed'
      severity: CRITICAL

  conditional:
    - id: bundle-size
      name: 'Bundle Size Impact'
      condition: 'frontend dependency'
      description: 'Bundle size change documented'
      severity: MEDIUM
```

### Refactor Evidence

```yaml
refactor_evidence:
  required:
    - id: behavior-preserved
      name: 'Behavior Preserved'
      description: 'Tests pass before and after refactor'
      severity: CRITICAL

    - id: no-new-features
      name: 'No New Features Added'
      description: 'Refactor is purely structural'
      severity: HIGH

  optional:
    - id: performance-improvement
      name: 'Performance Improvement'
      description: 'Benchmarks showing improvement if claimed'
      severity: LOW
```

---

## Workflow

### Phase 1: Detect Issue Type

```yaml
detection:
  bug_fix:
    patterns:
      - story title contains "fix", "bug", "issue"
      - commit messages contain "fix", "bug"
      - linked issue is type "bug"

  feature:
    patterns:
      - story title contains "add", "implement", "create"
      - acceptance criteria present

  dependency_update:
    patterns:
      - changes in package.json
      - changes in package-lock.json
      - story mentions "update", "upgrade", "dependency"

  refactor:
    patterns:
      - story title contains "refactor", "cleanup", "reorganize"
      - no new features in acceptance criteria
```

### Phase 2: Generate Checklist

```yaml
generate:
  - Load appropriate evidence template
  - Evaluate conditional items
  - Create evidence_checklist.md in story qa folder
  - Return checklist for QA verification
```

### Phase 3: Verify Evidence

```yaml
verify:
  for_each: checklist_item
  actions:
    - Check if evidence exists
    - Validate evidence quality
    - Mark as complete or missing

  output:
    complete: boolean
    missing: string[]
    blocking: boolean (if any CRITICAL missing)
```

---

## Command

```
*evidence-check {story-id} [--type bug_fix|feature|dependency_update|refactor]
```

**Examples:**

```bash
*evidence-check 6.3
*evidence-check 6.3 --type bug_fix
```

---

## Evidence Checklist Template

```markdown
# Evidence Checklist: Story {storyId}

**Type:** {issue_type}
**Generated:** {timestamp}
**Reviewer:** Quinn (Test Architect)

---

## Required Evidence

### 1. {evidence_name}

**Status:** [ ] Complete / [ ] Missing

**Description:** {description}

**Evidence Location:**

<!-- Link to screenshot, test file, commit, or explanation -->

**Verified By:**

<!-- QA reviewer name and date -->

---

### 2. {evidence_name}

...

---

## Conditional Evidence

### {evidence_name} (if {condition})

**Applicable:** [ ] Yes / [ ] No

**Status:** [ ] Complete / [ ] Missing / [ ] N/A

---

## Summary

| Category | Required | Provided | Missing |
| -------- | -------- | -------- | ------- |
| Critical | {count}  | {count}  | {count} |
| High     | {count}  | {count}  | {count} |
| Medium   | {count}  | {count}  | {count} |

**Verdict:** {COMPLETE | INCOMPLETE}

**Missing Critical Evidence:**
{list of missing critical items}

---

_Generated by Quinn (@qa) - Evidence Requirements Task_
```

---

## Integration with QA Review

```yaml
integration:
  trigger: During *review-build Phase 8 (Report Generation)

  workflow: 1. Detect issue type from story/commits
    2. Generate evidence checklist
    3. Verify each evidence item
    4. Include in qa_report.md
    5. Block if CRITICAL evidence missing

  blocking_rules:
    - Any CRITICAL evidence missing → REJECT
    - 2+ HIGH evidence missing → REJECT
    - Only MEDIUM/LOW missing → APPROVE with notes
```

---

## Metadata

```yaml
metadata:
  version: 1.0.0
  source: Auto-Claude PR Review Phase 3
  tags:
    - qa-enhancement
    - evidence-based
    - quality-gate
  updated_at: 2026-01-29
```