prd-completeness-auditor

What this skill does

# PRD Completeness Auditor

A systematic audit system that evaluates Product Requirements Documents against a comprehensive 116-item checklist, detecting gaps, classifying severity, identifying anti-patterns, and generating actionable remediation plans. Implements the GAP-AUDIT pattern (PATTERN-08) to produce CONTRACT-07 compliant GAP-INVENTORY output.

---

## 1. Purpose

This skill provides 12 core capabilities:

| # | Capability | Phase | Description |
|---|------------|-------|-------------|
| 1 | **Reference Load** | 1 | Load PRD checklist and calibrate to document type |
| 2 | **Document Parse** | 1 | Extract PRD structure and section boundaries |
| 3 | **MECE Map** | 2 | Apply 10-dimension coverage framework to PRD sections |
| 4 | **Item-by-Item Verify** | 3 | Check each of 116 checklist items against PRD content |
| 5 | **Gap Detect** | 3 | Identify missing, incomplete, or ambiguous elements |
| 6 | **Classify Gap Type** | 4 | Categorize gaps using 6-type taxonomy |
| 7 | **Score Severity** | 4 | Apply RUBRIC-07 (impact, likelihood, detectability) |
| 8 | **Anti-Pattern Scan** | 4 | Detect 15+ known PRD anti-patterns |
| 9 | **Evidence Cite** | 4 | Link each gap to specific PRD location with evidence |
| 10 | **Impact Assess** | 5 | Evaluate downstream development impact per gap |
| 11 | **Remediate Suggest** | 5 | Generate fix recommendations from pattern catalog |
| 12 | **Synthesize Report** | 6 | Produce GAP-INVENTORY + executive summary + remediation plan |

---

## 2. When to Use

**Ideal for:**
- Reviewing PRDs before development handoff
- Assessing requirements completeness for sprint planning
- Validating PRD against organizational standards
- Pre-flight check before stakeholder sign-off
- Identifying gaps that could cause scope creep
- Auditing legacy PRDs for maintenance or migration
- Quality gate before technical design phase

**Avoid when:**
- PRD is in early ideation phase (too early for formal audit)
- Document is not a PRD (use appropriate auditor)
- PRD author is unavailable for clarification on gaps
- Quick brainstorm or exploration (formal audit adds overhead)
- Document is a technical spec, not product requirements

---

## 3. Parameters

| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| `prd_content` | string | **yes** | — | PRD document content or file reference |
| `prd_type` | enum | no | `feature` | feature \| epic \| product \| initiative |
| `audit_depth` | enum | no | `standard` | quick \| standard \| comprehensive |
| `severity_threshold` | enum | no | `all` | critical_only \| high_and_above \| all |
| `focus_areas` | list | no | `all` | Specific sections to emphasize in audit |
| `include_anti_patterns` | boolean | no | `true` | Whether to scan for PRD anti-patterns |
| `output_format` | enum | no | `full` | full \| executive \| remediation_only |

### PRD Type Effects

| Type | Checklist Items | Focus Areas |
|------|-----------------|-------------|
| **feature** | 106 items | User stories, acceptance criteria, functional requirements |
| **epic** | 106 items | Cross-feature dependencies, phasing, rollout strategy |
| **product** | 100 items | Strategic alignment, market context, success metrics |
| **initiative** | 92 items | Business case, organizational impact, stakeholders |

### Audit Depth Effects

| Depth | Behavior |
|-------|----------|
| **quick** | High-priority items only (40 items), structural gaps, no anti-patterns |
| **standard** | Full checklist, anti-patterns, standard remediation |
| **comprehensive** | Full checklist, deep anti-pattern analysis, detailed remediation with templates |

---

## 4. Six-Phase Workflow

### Phase 1: Reference Loading & Context

**Purpose:** Load checklist and establish audit context.

**Steps:**
1. Receive PRD content (inline or file reference)
2. Load `prd-requirements-checklist.md` reference
3. Calibrate checklist based on `prd_type`:
   - Filter items by applicability matrix
   - Adjust severity levels for document type
   - Set critical items that must be present
4. Parse PRD to identify structure:
   - Extract section headers and hierarchy
   - Identify section boundaries
   - Note document format (markdown, docx, wiki)
5. Load `mece-coverage-framework.md` for structural validation
6. Initialize gap tracking structure with empty findings

**Techniques Used:**
- `document_structure_extraction` - Parse document hierarchy
- `checklist_calibration` - Adjust criteria for context

**Quality Gate:** Checklist calibrated for PRD type; PRD structure parsed with sections identified

**Output:** Calibrated checklist (N items) + PRD structure map

---

### Phase 2: MECE Scope Mapping

**Purpose:** Map PRD sections to coverage dimensions and identify structural gaps.

**Steps:**
1. Load 10 MECE coverage dimensions from `mece-coverage-framework.md`:
   - D1: Problem Space
   - D2: User Context
   - D3: Business Context
   - D4: Solution Requirements
   - D5: User Experience
   - D6: Quality Criteria
   - D7: Dependencies
   - D8: Risk Landscape
   - D9: Execution Context
   - D10: Technical Fit
2. Map each PRD section to exactly one dimension:
   - Assign based on content, not just header
   - Flag orphan sections (don't fit any dimension)
   - Flag missing dimensions (no sections mapped)
3. Identify structural gaps:
   - Dimension with no coverage = CRITICAL structural gap
   - Dimension with partial coverage = SIGNIFICANT structural gap
4. Note sections present but outside standard structure
5. Calculate initial coverage score per dimension

**Techniques Used:**
- `mece_gap_detection` (CAT-PR-GA) - Identify coverage gaps in structure
- `completeness_verification` (CAT-PR-GA) - Verify all dimensions addressed

**Quality Gate:** All 10 MECE dimensions assessed; structural gaps flagged

**Output:** Coverage mapping with dimension scores and structural gap list

---

### Phase 3: Item-by-Item Verification

**Purpose:** Verify each checklist item against PRD content.

**Steps:**

For each checklist item (PC-01 through TC-08):

```
1. LOCATE RELEVANT SECTION
   └─ Based on dimension mapping from Phase 2

2. ASSESS PRESENCE
   ├─ PRESENT: Item clearly addressed
   ├─ PARTIAL: Item mentioned but incomplete
   └─ ABSENT: No evidence of item

3. EVALUATE QUALITY (if present)
   ├─ CLEAR: Unambiguous and actionable
   ├─ AMBIGUOUS: Multiple interpretations possible
   ├─ INCONSISTENT: Contradicts other content
   ├─ INCOMPLETE: Missing key details
   └─ OUTDATED: Stale information

4. RECORD EVIDENCE
   ├─ Location: Section/line reference
   ├─ Quote: Supporting text from PRD
   └─ Assessment: Pass/fail with rationale

5. FLAG FOR GAP CLASSIFICATION (if needed)
   └─ Any non-PRESENT or non-CLEAR items
```

**Verification Rules by Section:**

| Section | Items | Critical Items | Verification Approach |
|---------|-------|----------------|----------------------|
| Problem & Context | PC-01 to PC-15 | PC-01, PC-02, PC-05 | Check for evidence-backed problem |
| User & Stakeholder | US-01 to US-12 | US-01, US-02, US-05 | Verify persona specificity |
| Goals & Success | GS-01 to GS-10 | GS-01, GS-02, GS-03 | Check measurability |
| Requirements | RQ-01 to RQ-18 | RQ-01, RQ-02, RQ-05 | Verify testability |
| User Stories | ST-01 to ST-14 | ST-01, ST-02 | Check INVEST compliance |
| Acceptance Criteria | AC-01 to AC-12 | AC-01, AC-02, AC-03 | Verify Gherkin-like structure |
| Dependencies | DP-01 to DP-08 | DP-01, DP-02 | Check bidirectional links |
| Risks & Assumptions | RA-01 to RA-10 | RA-01, RA-02 | Verify mitigation plans |
| Timeline & Scope | TS-01 to TS-09 | TS-01, TS-04, TS-05 | Check explicit boundaries |
| Technical | TC-01 to TC-08 | TC-01 (if applicable) | Verify feasibility notes |

**Techniques Used:**
- `completeness_verification` (CAT-PR-GA) - Check presence of each item
- `negative_space_analysis` (CAT-PR-GA) - Identify what's NOT said
- `exhaustive_edge_case_enumeration` (CAT-PR-CE) - Find uncovered sce