Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计提醒

doc-prd-reviewer文档 PRD 审稿人

Agent Skill

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

总安装

1,721

周安装

71

GitHub Stars

14

下载量

562
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:doc-prd-reviewer(文档 PRD 审稿人)
来源仓库:https://github.com/vladm3105/aidoc-flow-framework
仓库路径:skills/doc-prd-reviewer
安装命令:
npx skills add https://github.com/vladm3105/aidoc-flow-framework --skill doc-prd-reviewer
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/vladm3105/aidoc-flow-framework --skill doc-prd-reviewer

简介

深度审查 PRD 内容质量与逻辑一致性。

  • 检查链接有效性、BRD 对齐情况及可读性。
  • 标记需人工介入的复杂问题如策略分歧。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • 已并入 doc-prd-audit,仅作历史参考保留。
  • doc-prd-reviewer 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

doc-prd-reviewer

DEPRECATED (2026-03-02): This skill has been merged into doc-prd-audit as part of the 2-skill model standardization. Use doc-prd-audit for all PRD review needs. This file is retained for reference only. Replacement: /doc-prd-audit Migration: No action required. Simply use doc-prd-audit instead of this skill.

Legacy Documentation (Reference Only)

Purpose

Comprehensive content review and quality assurance for Product Requirements Documents (PRD). This skill performs deep content analysis beyond structural validation, checking link integrity, threshold consistency, BRD alignment, and identifying issues that require manual business review.

Layer: 2 (PRD Quality Assurance)

Upstream: PRD (from doc-prd-autopilot or doc-prd)

Downstream: None (final QA gate before EARS generation)


When to Use This Skill

Use doc-prd-reviewer when:

  • After PRD Generation: Run immediately after doc-prd-autopilot completes
  • Manual PRD Edits: After making manual changes to a PRD
  • Pre-EARS Check: Before running doc-ears-autopilot
  • Periodic Review: Regular quality checks on existing PRDs
  • CI/CD Integration: Automated review gate in documentation pipelines

Do NOT use when:

  • PRD does not exist yet (use doc-prd or doc-prd-autopilot first)
  • Need structural/schema validation only (use doc-prd-validator)
  • Generating new PRD content (use doc-prd)

Skill vs Validator: Key Differences

Aspectdoc-prd-validatordoc-prd-reviewer
FocusSchema compliance, EARS-Ready scoreContent quality, cross-document consistency
ChecksRequired sections, field formatsLink integrity, threshold alignment, placeholders
Auto-FixStructural issues onlyContent issues (links, dates, placeholders)
OutputEARS-Ready score (numeric)Review score + issue list
PhasePhase 4 (Validation)Phase 5 (Final Review)
BlockingEARS-Ready < threshold blocksReview score < threshold flags

Review Workflow

flowchart TD
    A[Input: PRD Path] --> B[Load PRD Files]
    B --> C{Sectioned or Monolithic?}

    C -->|Sectioned| D[Load All Section Files]
    C -->|Monolithic| E[Load Single File]

    D --> F[Run Review Checks]
    E --> F

    subgraph Review["Review Checks"]
        F --> G[1. Link Integrity]
        G --> H[2. Threshold Consistency]
      H --> H2[2a. Diagram Contract Compliance]
      H2 --> I[3. BRD Alignment]
        I --> J[4. Placeholder Detection]
        J --> K[5. Traceability Tags]
        K --> L[6. Section Completeness]
        L --> M[7. Customer Content]
        M --> M2[8. Naming Compliance]
        M2 --> M3[9. Structure Compliance]
        M3 --> M4[10. Upstream Drift Detection]
    end

    M4 --> N{Issues Found?}
    N -->|Yes| O[Categorize Issues]
    O --> P{Auto-Fixable?}
    P -->|Yes| Q[Apply Auto-Fixes]
    Q --> R[Re-run Affected Checks]
    P -->|No| S[Flag for Manual Review]
    R --> N
    S --> T[Generate Report]
    N -->|No| T
    T --> U[Calculate Review Score]
    U --> V{Score >= Threshold?}
    V -->|Yes| W[PASS]
    V -->|No| X[FAIL with Details]

Review Checks

1. Link Integrity

Validates all internal document links resolve correctly.

Scope:

  • Navigation links ([Previous:...], [Next:...])
  • Section cross-references ([See Section 8.2](...))
  • Index to section links
  • External documentation links (warns if unreachable)

Detection:

Scanning PRD-01 for internal links...
├── PRD-01.md: 17 links found
│   ├── PRD-01.1_document_control.md ✓
│   ├── PRD-01.2_executive_summary.md ✓
│   ├── PRD-01.18_missing_section.md ✗ NOT FOUND
│   └── ...
└── Result: 1 broken link found

Auto-Fix:

  • Remove broken links to non-existent sections
  • Correct path typos if target exists with similar name
  • Update relative paths if file was moved

Error Codes:

CodeSeverityDescription
REV-L001ErrorBroken internal link
REV-L002WarningExternal link unreachable
REV-L003InfoLink path uses absolute instead of relative

2. Threshold Consistency

Verifies performance metrics match across all sections.

Scope:

  • Section 5 (Success Metrics) thresholds
  • Section 9 (Quality Attributes) performance targets
  • Section 17/20 (Appendix B.1) timing profile matrix
  • Cross-references to BRD thresholds

Detection:

Comparing thresholds across sections...
├── Operation: auth.p99
│   ├── Section 5 (Success Metrics): 100ms ✓
│   ├── Section 9 (Quality Attributes): 100ms ✓
│   ├── Section 17 (Timing Matrix): 100ms ✓
│   └── BRD-01 (Source): 100ms ✓
├── Operation: authz.p99
│   ├── Section 5: 10ms ✓
│   ├── Section 9: 15ms ✗ MISMATCH
│   ├── Section 17: 10ms ✓
│   └── BRD-01: 10ms ✓
└── Result: 1 inconsistency found

Auto-Fix:

  • Align PRD values to BRD source (authoritative)
  • Update all sections to use consistent value
  • Add comment noting auto-fix with timestamp

Error Codes:

CodeSeverityDescription
REV-T001ErrorThreshold mismatch across PRD sections
REV-T002ErrorThreshold differs from BRD source
REV-T003WarningThreshold unit inconsistency (ms vs s)
REV-T004InfoThreshold stricter than BRD (acceptable)

2a. Diagram Contract Compliance

Validates PRD diagram contract requirements defined by ai_dev_ssd_flow/DIAGRAM_STANDARDS.md.

Scope:

  • Required PRD tags: @diagram: c4-l2, @diagram: dfd-l1, @diagram: sequence-*
  • Required sequence error/alternate path coverage for critical journeys
  • Intent header fields: diagram_type, level, scope_boundary, upstream_refs, downstream_refs

Error Codes:

CodeSeverityDescription
REV-DC001ErrorMissing required PRD diagram tag (c4-l2, dfd-l1, or sequence-*)
REV-DC002ErrorRequired sequence flow missing explicit exception path
REV-DC003WarningDiagram intent header missing required fields

3. BRD Alignment

Validates PRD requirements accurately reflect BRD source.

Scope:

  • Every PRD requirement maps to a BRD requirement
  • No orphaned requirements (PRD without BRD source)
  • No missing requirements (BRD without PRD mapping)
  • Scope boundaries match

Detection:

Verifying PRD requirements map to BRD source...
├── PRD.01.0101 → BRD.01.0101 (Multi-Provider Auth) ✓
├── PRD.01.0102 → BRD.01.0102 (4D Authorization) ✓
├── PRD.01.0103 → BRD.01.0103 (Trust Levels) ✓
├── PRD.01.0109 → ??? ✗ NO BRD MAPPING
├── BRD.01.0110 → ??? ✗ NO PRD MAPPING (deferred)
└── Result: 11/12 aligned, 1 orphan, 1 missing

Scope Alignment:
├── PRD In-Scope vs BRD In-Scope: MATCH ✓
├── PRD Out-Scope vs BRD Out-Scope: MATCH ✓
└── Deferred items: 2 (correctly marked)

Manual Review Required:

  • Orphaned requirements need justification
  • Missing requirements need explicit deferral note
  • Scope mismatches need business decision

Error Codes:

CodeSeverityDescription
REV-A001ErrorPRD requirement without BRD source
REV-A002WarningBRD requirement without PRD mapping
REV-A003ErrorScope mismatch (PRD vs BRD)
REV-A004InfoRequirement correctly marked as deferred

4. Placeholder Detection

Identifies incomplete content requiring replacement.

Scope:

  • [TODO], [TBD], [PLACEHOLDER] text
  • Template dates: YYYY-MM-DDTHH:MM:SS, MM/DD/YYYY
  • Template names: [Name], [Author], [Reviewer]
  • Empty sections: <!-- Content here -->
  • Lorem ipsum or sample text

Detection:

Scanning for placeholder text...
├── [TODO]: 0 found ✓
├── [TBD]: 1 found
│   └── PRD-01.10_customer_content.md:42: "[TBD: Marketing copy]"
├── YYYY-MM-DDTHH:MM:SS: 0 found ✓
├── [Name]: 0 found ✓
├── Empty sections: 0 found ✓
└── Result: 1 placeholder found

Auto-Fix:

  • Replace YYYY-MM-DDTHH:MM:SS with current datetime
  • Replace [Name] with document author from metadata
  • Remove empty comment placeholders
  • Flag [TODO]/[TBD] for manual completion

Error Codes:

CodeSeverityDescription
REV-P001Error[TODO] placeholder found
REV-P002Error[TBD] placeholder found
REV-P003WarningTemplate date not replaced
REV-P004WarningTemplate name not replaced
REV-P005WarningEmpty section content

5. Traceability Tags

Validates @brd:, @depends:, and @discoverability: tags.

Scope:

  • All @brd: BRD.XX.XX.XX tags reference valid BRD IDs
  • @depends: PRD-XX references exist
  • @discoverability: PRD-XX references exist
  • Tag format follows convention

Detection:

Validating traceability tags...
├── @brd tags: 15 found
│   ├── BRD.01.0101 ✓ (exists in BRD-01)
│   ├── BRD.01.0102 ✓
│   ├── BRD.01.9901 ✗ NOT FOUND in BRD
│   └── ...
├── @depends tags: 2 found
│   ├── PRD-06 ✓ (exists)
│   └── PRD-07 ✓ (exists)
├── @discoverability tags: 2 found
│   ├── PRD-02 ✗ (planned, not yet created)
│   └── PRD-03 ✗ (planned, not yet created)
└── Result: 1 invalid @brd, 2 forward references (acceptable)

Auto-Fix:

  • Remove invalid @brd tags
  • Correct typos in tag IDs if close match exists
  • Add (planned) suffix to forward references

Error Codes:

CodeSeverityDescription
REV-TR001Error@brd tag references non-existent BRD ID
REV-TR002Warning@depends references non-existent PRD
REV-TR003Info@discoverability is forward reference
REV-TR004WarningTag format malformed

6. Section Completeness

Verifies all required sections have substantive content.

Scope:

  • Minimum word count per section (configurable)
  • Section headers present
  • Tables have data rows (not just headers)
  • Mermaid diagrams render properly

Detection:

Checking section completeness...
├── Section 1 (Document Control): 150 words ✓
├── Section 2 (Executive Summary): 280 words ✓
├── Section 10 (Customer Content): 45 words ⚠ THIN
├── Section 17 (Appendices): 820 words ✓
├── Table check: 12 tables, all have data rows ✓
├── Mermaid check: 3 diagrams, all valid ✓
└── Result: 16/17 complete, 1 thin section

Minimum Word Counts (configurable):

SectionMinimum Words
Executive Summary100
Problem Statement75
Functional Requirements200
Quality Attributes100
Customer Content100
Risk Assessment75
Appendices200

Error Codes:

CodeSeverityDescription
REV-S001ErrorRequired section missing entirely
REV-S002WarningSection below minimum word count
REV-S003WarningTable has no data rows
REV-S004ErrorMermaid diagram syntax error

7. Customer Content Review

Flags Section 10 for mandatory business/marketing review.

Scope:

  • Section 10 (Customer-Facing Content) exists
  • Content is substantive (not placeholder)
  • Marketing language appropriate
  • No technical jargon in customer-facing text

Detection:

Reviewing customer-facing content (Section 10)...
├── Section exists: ✓
├── Word count: 180 words ✓
├── Placeholder check: PASS ✓
├── Technical jargon scan:
│   ├── "4D Matrix" - technical term found
│   ├── "OIDC" - acronym may need explanation
│   └── "mTLS" - technical term found
└── Result: FLAG for marketing review (3 technical terms)

Manual Review Required:

  • All Section 10 content requires business sign-off
  • Technical terms need customer-friendly alternatives
  • Messaging alignment with brand guidelines

Error Codes:

CodeSeverityDescription
REV-C001ErrorSection 10 missing
REV-C002WarningSection 10 is placeholder content
REV-C003InfoTechnical jargon in customer content
REV-C004FlagRequires marketing/business review

8. Naming Compliance

Validates element IDs and threshold tags follow doc-naming standards.

Scope:

  • Element IDs use PRD.NN.xxxx format
  • Element type codes valid for PRD (01-09, 11, 22, 24)
  • Threshold tags use @threshold: PRD.NN.key format
  • No legacy patterns (US-NNN, FR-NNN, AC-NNN, F-NNN)

Detection:

Validating naming compliance (per doc-naming skill)...
├── Element IDs: 24 found
│   ├── PRD.01.0101 ✓ (valid format, code 01 valid for PRD)
│   ├── PRD.01.0905 ✓ (valid format, code 09 valid for PRD)
│   ├── PRD.01.2501 ✗ (code 25 not valid for PRD - EARS only)
│   └── ...
├── Threshold tags: 8 found
│   ├── @threshold: PRD.01.perf.auth.p99 ✓
│   ├── @threshold: perf.auth.p99 ✗ (missing PRD.NN prefix)
│   └── ...
├── Legacy patterns: 2 found
│   ├── US-001 ✗ (deprecated - use PRD.NN.09.SS)
│   └── FR-003 ✗ (deprecated - use PRD.NN.01.SS)
└── Result: 3 naming violations

Auto-Fix:

  • Convert legacy patterns to unified format
  • Add missing PRD.NN prefix to threshold tags
  • Suggest correct element type codes

Error Codes:

CodeSeverityDescription
REV-N001ErrorInvalid element ID format
REV-N002ErrorElement type code not valid for PRD
REV-N003ErrorLegacy pattern detected (US-NNN, FR-NNN, etc.)
REV-N004ErrorThreshold tag missing document reference
REV-N005WarningThreshold key format non-standard

Reference: See doc-naming skill for complete naming rules.


9. Structure Compliance (CRITICAL)

Validates PRD follows the mandatory nested folder rule.

Nested Folder Rule: ALL PRDs MUST be in nested folders regardless of size.

Required Structure:

PRD TypeRequired Location
Monolithicdocs/02_PRD/PRD-NN_{slug}/PRD-NN_{slug}.md
Sectioneddocs/02_PRD/PRD-NN_{slug}/PRD-NN.0_index.md, PRD-NN.1_*.md, etc.

Detection:

Validating folder structure...
├── PRD Location: docs/02_PRD/PRD-01_f1_iam/PRD-01_f1_iam.md
├── Expected Folder: PRD-01_f1_iam ✓
├── Parent Path: docs/02_PRD/ ✓
├── Nested Structure: Valid ✓
└── Result: PASS

Failure Detection:

Validating folder structure...
├── PRD Location: docs/02_PRD/PRD-02_f2_session.md
├── Expected: docs/02_PRD/PRD-02_f2_session/PRD-02_f2_session.md
├── Nested Structure: INVALID ✗
└── Result: FAIL - PRD not in nested folder

Auto-Fix:

  1. Create the nested folder docs/02_PRD/PRD-NN_{slug}/
  2. Move the PRD file(s) into the folder
  3. Update all internal links (navigation, cross-references)
  4. Update upstream BRD links: ../01_BRD/../../01_BRD/
  5. Update any glossary links to correct path

Error Codes:

CodeSeverityDescription
REV-STR001ErrorPRD not in nested folder (blocking)
REV-STR002ErrorPRD folder name doesn't match PRD ID
REV-STR003WarningMonolithic PRD should be sectioned (>25KB)
REV-STR004ErrorBRD link path incorrect for nested folder

This check is BLOCKING - PRD must pass structure validation before other checks proceed.


10. Upstream Drift Detection (Mandatory Cache)

Detects when upstream source documents have been modified after the PRD was created or last updated. The drift cache is mandatory - the reviewer MUST create/update it after every review.

Purpose: Identifies stale PRD content that may not reflect current BRD documentation. When upstream documents (BRD specifications, business requirements) change, the PRD may need updates to maintain alignment.

Scope:

  • @ref: tag targets (BRD documents, technical specifications)
  • @brd: tag references
  • Traceability section upstream artifact links
  • Any markdown links to ../01_BRD/ or source documents

10.1 Drift Cache File (MANDATORY)

Location: docs/02_PRD/{PRD_folder}/.drift_cache.json

IMPORTANT: The reviewer MUST:

  1. Read the cache if it exists (for hash comparison)
  2. Create the cache if it doesn't exist
  3. Update the cache after every review with current hashes

Cache Schema:

{
  "schema_version": "1.0",
  "document_id": "PRD-01",
  "document_version": "1.0",
  "last_reviewed": "2026-02-10T17:00:00",
  "reviewer_version": "1.4",
  "upstream_documents": {
    "../../01_BRD/BRD-01_f1_iam/BRD-01.md": {
      "hash": "sha256:a1b2c3d4e5f6g7h8i9j0...",
      "last_modified": "2026-02-10T15:34:26",
      "file_size": 50781,
      "version": "1.0",
      "sections_tracked": ["#7-functional-requirements", "#8-non-functional-requirements"]
    },
    "../../01_BRD/BRD-01_f1_iam/BRD-01.7_functional_requirements.md": {
      "hash": "sha256:k1l2m3n4o5p6q7r8s9t0...",
      "last_modified": "2026-02-10T15:34:21",
      "file_size": 4730,
      "version": "1.0",
      "sections_tracked": ["#authentication", "#authorization"]
    }
  },
  "review_history": [
    {
      "date": "2026-02-10T16:30:00",
      "score": 97,
      "drift_detected": false,
      "report_version": "v002"
    }
  ]
}

10.2 Detection Algorithm (Three-Phase)

PHASE 1: Load Cache (if exists)
=========================================
1. Check for .drift_cache.json in PRD folder
2. If exists:
   - Load cached hashes and metadata
   - Set detection_mode = "hash_comparison"
3. If not exists:
   - Set detection_mode = "timestamp_only"
   - Will create cache at end of review

PHASE 2: Detect Drift
=========================================
For each upstream reference in PRD:

  A. Extract reference:
     - @ref: tags → [path, section anchor]
     - @brd: tags → [BRD ID, requirement ID]
     - Links to ../01_BRD/ → [path]
     - Traceability table upstream artifacts → [path]

  B. Resolve and validate:
     - Resolve path to absolute file path
     - Check file exists (skip if covered by Check #1)
     - Get file stats: mtime, size

  C. Compare (based on detection_mode):

     IF detection_mode == "hash_comparison":
       - Compute SHA-256 hash of current file content
       - Compare to cached hash
       - IF hash differs:
           - Calculate change_percentage
           - Flag as CONTENT_DRIFT (REV-D002)
           - IF change > 20%: Flag as CRITICAL (REV-D005)

     ELSE (timestamp_only):
       - Compare file mtime > PRD last_updated
       - IF mtime > PRD date:
           - Flag as TIMESTAMP_DRIFT (REV-D001)

  D. Check version field (if YAML frontmatter):
     - Extract version from upstream doc
     - Compare to cached version
     - IF version incremented: Flag REV-D003

PHASE 3: Update Cache (MANDATORY)
=========================================
1. Compute SHA-256 hash for ALL upstream documents
2. Create/update .drift_cache.json with:
   - Current hashes
   - Current timestamps
   - Current file sizes
   - Review metadata
3. Append to review_history array

10.3 Hash Calculation (MANDATORY BASH EXECUTION)

CRITICAL: You MUST execute actual bash commands to compute hashes. DO NOT write placeholder values like verified_no_drift or pending_verification.

10.3.1 Compute File Hash

Execute this bash command for each upstream document:

sha256sum <file_path> | cut -d' ' -f1

Store result in drift cache as: "hash": "sha256:<64_hex_characters>"

10.3.2 Hash Format Validation

CheckRequirement
PrefixMust be sha256:
LengthExactly 64 hex characters after prefix
Characters[0-9a-f] only

REJECTED VALUES (re-compute immediately if found):

  • sha256:verified_no_drift
  • sha256:pending_verification
  • pending_verification
  • sha256:TBD
  • Any value where hex portion!= 64 characters

10.3.3 Verification After Cache Write

# Verify all hashes in cache are valid
grep -oP '"hash":\s*"sha256:[0-9a-f]{64}"' .drift_cache.json
# Must return valid matches for all upstream docs

10.3.4 Section Hash Computation

For anchor-specific tracking, extract section content first then hash:

# Extract section and compute hash
sed -n '/^## Section Name/,/^## /p' <file_path> | head -n -1 | sha256sum | cut -d' ' -f1

10.4 Error Codes

CodeSeverityDescriptionTrigger
REV-D001WarningUpstream document modified after PRD creationmtime > PRD date (no cache)
REV-D002WarningReferenced content has changedhash mismatch (with cache)
REV-D003InfoUpstream document version incrementedversion field changed
REV-D004InfoNew content added to upstreamfile size increased >10%
REV-D005ErrorCritical modification (>20% change)hash diff >20%
REV-D006InfoCache created (first review)no prior cache existed
REV-D009ErrorInvalid hash placeholder detectedhash is verified_no_drift, pending_verification, or invalid format

10.5 Report Output

## 9. Upstream Drift Detection (5/5)

### Cache Status

| Field | Value |
|-------|-------|
| Cache File | `.drift_cache.json` |
| Cache Status | ✅ Updated |
| Detection Mode | Hash Comparison |
| Documents Tracked | 2 |

### Upstream Document Analysis

| Upstream Document | Hash Status | Last Modified | Change % | Status |
|-------------------|-------------|---------------|----------|--------|
| BRD-01.md | ✅ Match | 2026-02-10T15:34:26 | 0% | Current |
| BRD-01.7_functional_requirements.md | ✅ Match | 2026-02-10T15:34:21 | 0% | Current |

### Drift Summary

| Status | Count | Details |
|--------|-------|---------|
| ✅ Current | 2 | All upstream documents synchronized |
| ⚠️ Warning | 0 | No drift detected |
| ❌ Critical | 0 | No major changes |

**Cache updated**: 2026-02-10T17:00:00

10.6 Configuration

SettingDefaultDescription
cache_enabledtrueMandatory - always create/update cache
drift_threshold_days7Days before timestamp drift becomes Warning
critical_change_pct20Percentage change for critical drift
track_sectionstrueTrack individual section hashes for anchored refs
max_history_entries10Maximum review_history entries to retain

Review Score Calculation

Scoring Formula:

CategoryWeightCalculation
Link Integrity10%(valid_links / total_links) × 10
Threshold Consistency10%(consistent_thresholds / total_thresholds) × 10
BRD Alignment18%(aligned_requirements / total_requirements) × 18
Placeholder Detection10%(no_placeholders? 10: 10 - (placeholder_count × 2))
Traceability Tags10%(valid_tags / total_tags) × 10
Section Completeness10%(complete_sections / total_sections) × 10
Customer Content5%(exists && substantive? 5: 0)
Naming Compliance10%(valid_ids / total_ids) × 10
Structure Compliance12%(nested_folder_valid? 12: 0) - BLOCKING
Upstream Drift5%(fresh_refs / total_refs) × 5

Total: Sum of all categories (max 100)

Note: Structure Compliance is a blocking check. If structure validation fails (REV-STR001), the review cannot pass regardless of other scores.

Thresholds:

  • PASS: ≥ 90 (configurable)
  • WARNING: 80-89
  • FAIL: < 80

Command Usage

Basic Usage

# Review specific PRD
/doc-prd-reviewer PRD-01

# Review PRD by path
/doc-prd-reviewer docs/02_PRD/PRD-01_f1_iam/

# Review all PRDs
/doc-prd-reviewer all

Options

OptionDefaultDescription
--min-score90Minimum passing review score
--auto-fixtrueApply automatic fixes
--no-auto-fixfalseDisable auto-fix (report only)
--checkallSpecific checks to run (comma-separated)
--skipnoneChecks to skip (comma-separated)
--verbosefalseDetailed output per check
--reporttrueGenerate markdown report
--report-pathautoReport location (auto = alongside document)
--versionautoReview version (auto = auto-increment)

Examples

# Review with verbose output
/doc-prd-reviewer PRD-01 --verbose

# Report only (no auto-fix)
/doc-prd-reviewer PRD-01 --no-auto-fix

# Run only specific checks
/doc-prd-reviewer PRD-01 --check links,placeholders

# Skip customer content check
/doc-prd-reviewer PRD-01 --skip customer

# Lower threshold for draft review
/doc-prd-reviewer PRD-01 --min-score 80

Configuration

# config/prd_reviewer.yaml
prd_reviewer:
  version: "1.0"

  scoring:
    min_score: 90
    weights:
      link_integrity: 15
      threshold_consistency: 20
      brd_alignment: 25
      placeholder_detection: 15
      traceability_tags: 10
      section_completeness: 10
      customer_content: 5

  checks:
    links:
      enabled: true
      check_external: false
      auto_fix: true
    thresholds:
      enabled: true
      source_priority: brd  # brd or prd
      auto_fix: true
    alignment:
      enabled: true
      allow_orphans: false
      require_deferral_notes: true
    placeholders:
      enabled: true
      patterns:
        - "[TODO]"
        - "[TBD]"
        - "[PLACEHOLDER]"
        - "YYYY-MM-DDTHH:MM:SS"
        - "[Name]"
      auto_fix_dates: true
      auto_fix_names: true
    traceability:
      enabled: true
      allow_forward_refs: true
      auto_fix_typos: true
    completeness:
      enabled: true
      min_words:
        executive_summary: 100
        problem_statement: 75
        functional_requirements: 200
        quality_attributes: 100
        customer_content: 100
        risk_assessment: 75
        appendices: 200
    customer:
      enabled: true
      flag_always: true
      jargon_patterns:
        - "OIDC"
        - "mTLS"
        - "JWT"
        - "4D Matrix"

  output:
    report_format: markdown
    report_path: auto  # auto = alongside reviewed document
    versioning:
      enabled: true
      pattern: "_vNNN"  # e.g., _v001, _v002
      auto_increment: true
    include_line_numbers: true
    include_fix_diff: true
    include_delta_report: true  # Compare with previous version

Output Report

File Naming & Versioning

Pattern: PRD-NN.R_review_report_vNNN.md

Nested Folder Rule: ALL PRDs use nested folders (PRD-NN_{slug}/) regardless of size. This ensures review reports, fix reports, and drift cache files are organized with their parent document.

Location: Inside the PRD nested folder: docs/02_PRD/PRD-NN_{slug}/

Versioning Rules:

  1. First Review: Creates PRD-NN.R_review_report_v001.md
  2. Subsequent Reviews: Auto-increments version (v002, v003, etc.)
  3. Same-Day Reviews: Each review gets unique version number

Example Directory:

docs/02_PRD/PRD-01_f1_iam/
├── PRD-01.md
├── PRD-01.R_review_report_v001.md    # First review
├── PRD-01.R_review_report_v002.md    # After fixes
└── PRD-01.R_review_report_v003.md    # Final review

Report Structure

# PRD Review Report: PRD-01 (v003)

**Review Date**: 2026-02-08T14:30:00
**Review Version**: v003
**PRD**: PRD-01 (F1 Identity & Access Management)
**Status**: PASS
**Review Score**: 95/100

---

## Summary

| Check | Status | Score | Issues |
|-------|--------|-------|--------|
| Link Integrity | PASS | 15/15 | 0 |
| Threshold Consistency | PASS | 20/20 | 0 |
| BRD Alignment | PASS | 23/25 | 2 deferred |
| Placeholder Detection | PASS | 15/15 | 0 |
| Traceability Tags | PASS | 10/10 | 0 |
| Section Completeness | PASS | 10/10 | 0 |
| Customer Content | FLAG | 2/5 | Needs review |
| **Total** | **PASS** | **95/100** | |

---

## Issues Detail

### Flagged for Manual Review

| File | Line | Issue | Code |
|------|------|-------|------|
| PRD-01.10_customer_content.md | 42 | Technical jargon: "4D Matrix" | REV-C003 |
| PRD-01.10_customer_content.md | 58 | Technical jargon: "OIDC" | REV-C003 |
| PRD-01.10_customer_content.md | - | Requires marketing review | REV-C004 |

### Auto-Fixes Applied

| File | Line | Fix | Code |
|------|------|-----|------|
| PRD-01.1_document_control.md | 15 | Date: YYYY-MM-DDTHH:MM:SS → 2026-02-08T14:30:00 | REV-P003 |

### Deferred Items (Acceptable)

| Requirement | Reason | Note |
|-------------|--------|------|
| BRD.01.0110 | Device Trust | Marked P3, next cycle |
| BRD.01.0112 | Time-Based Access | Marked P3, next cycle |

---

## Recommendations

1. **Section 10 Review**: Schedule marketing review for customer-facing content
2. **Technical Terms**: Consider adding glossary link for "4D Matrix", "OIDC"
3. **Deferred Items**: Confirm P3 priority with stakeholders

---

**Generated By**: doc-prd-reviewer v1.1
**Report Location**: docs/02_PRD/PRD-01_f1_iam/PRD-01.R_review_report_v003.md
**Previous Review**: v002 (Score: 91/100)

Delta Reporting

When previous reviews exist, include score comparison:

## Score Comparison (v002 → v003)

| Metric | Previous (v002) | Current (v003) | Delta |
|--------|-----------------|----------------|-------|
| Overall Score | 91 | 95 | +4 |
| Errors | 2 | 0 | -2 |
| Warnings | 5 | 3 | -2 |
| Issues Resolved | - | 4 | - |
| New Issues | - | 0 | - |

See REVIEW_DOCUMENT_STANDARDS.md for complete versioning requirements.


Integration with doc-prd-autopilot

This skill is automatically invoked during Phase 5 of doc-prd-autopilot:

flowchart LR
    A[Phase 4: Validation] --> B[Phase 5: Final Review]
    B --> C{doc-prd-reviewer}
    C --> D[Phase 6: Continue]

Autopilot Integration:

# In doc-prd-autopilot config
review:
  enabled: true
  skill: doc-prd-reviewer
  options:
    min_score: 90
    auto_fix: true
    flag_customer_content: true

Standalone vs Integrated Use

Use CaseCommandNotes
After autopilotAutomaticInvoked as Phase 5
Manual PRD edit/doc-prd-reviewer PRD-01Run explicitly
Pre-EARS check/doc-prd-reviewer PRD-01Verify before EARS
Periodic audit/doc-prd-reviewer allReview all PRDs

Error Recovery

Common Issues

ErrorCauseResolution
PRDNotFoundErrorPRD path invalidCheck path or PRD ID
BRDNotFoundErrorBRD source missingVerify BRD exists for alignment check
LowScoreErrorScore below thresholdFix flagged issues

Recovery Commands

# Re-run after fixes
/doc-prd-reviewer PRD-01

# Run with lower threshold (for drafts)
/doc-prd-reviewer PRD-01 --min-score 75

# Skip failing check temporarily
/doc-prd-reviewer PRD-01 --skip alignment

Related Skills

SkillRelationship
doc-namingNaming standards for Check #8
doc-prd-autopilotInvokes this skill in Phase 5
doc-prd-validatorStructural validation (Phase 4)
doc-prd-fixerApplies fixes based on review findings
doc-prd-auditWrapper that combines this reviewer output with validator findings
doc-prdPRD creation rules
doc-brd-auditBRD source validation (unified quality gate)
doc-ears-autopilotDownstream consumer

Version History

VersionDateChanges
2.02026-03-02DEPRECATED: Merged into doc-prd-audit as part of 2-skill model standardization. Use doc-prd-audit instead.
1.92026-03-012-Skill BRD Model: Updated BRD validation references from doc-brd-validator to doc-brd-audit (unified quality gate)
1.82026-02-27Hash Computation Fix: Replaced Python pseudocode with mandatory bash sha256sum execution; Added hash validation (REV-D009); Reject placeholder values (verified_no_drift, pending_verification)
1.72026-02-26Migrated frontmatter to metadata schema; documented relationship with doc-prd-audit wrapper
1.62026-02-11Numbering Fix: Corrected Check #10 subsections from 9.1-9.6 to 10.1-10.6 (drift cache, detection algorithm, hash calculation, error codes, report output, configuration)
1.52026-02-11Structure Compliance: Added Check #9 for nested folder rule enforcement (REV-STR001-STR004); Shifted Upstream Drift to Check #10; Added structure compliance to scoring (12% weight, blocking); Updated workflow diagram to 10 checks
1.42026-02-10T17:00:00Mandatory drift cache: Reviewer MUST create/update .drift_cache.json after every review; Three-phase detection algorithm; SHA-256 hash computation; Hash comparison mode when cache exists; REV-D006 code for cache creation; Cache schema with review_history tracking
1.32026-02-10Added Check #9: Upstream Drift Detection - detects when BRD documents modified after PRD creation; REV-D001-D005 error codes; drift configuration; Added doc-prd-fixer to related skills
1.22026-02-10Added review versioning support (_vNNN pattern); Delta reporting for score comparison
1.12026-02-08Added Check #8: Naming Compliance (doc-naming integration)
1.02026-02-08Initial skill creation with 7 review checks

Implementation Plan Consistency (IPLAN-004)

  • Treat plan-derived outputs as valid source mode and verify intent preservation from implementation plan scope/objectives.
  • Validate upstream autopilot precedence assumption: --iplan > --ref > --prompt.
  • Flag objective/scope conflicts between plan context and artifact output as blocking issues requiring clarification.
  • Do not introduce legacy fallback paths such as docs-v2.0/00_REF.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.99%
按下载量换算208

Claude

30.06%
按下载量换算169

Cursor

19.65%
按下载量换算110

Gemini CLI

11.07%
按下载量换算62

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills