Nelze vybrat více než 25 témat Téma musí začínat písmenem nebo číslem, může obsahovat pomlčky („-“) a může být dlouhé až 35 znaků.

step-v-11-holistic-quality-validation.md 8.5KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261
  1. ---
  2. # File references (ONLY variables used in this step)
  3. nextStepFile: './step-v-12-completeness-validation.md'
  4. prdFile: '{prd_file_path}'
  5. validationReportPath: '{validation_report_path}'
  6. ---
  7. # Step 11: Holistic Quality Assessment
  8. ## STEP GOAL:
  9. Assess the PRD as a cohesive, compelling document - evaluating document flow, dual audience effectiveness (humans and LLMs), BMAD PRD principles compliance, and overall quality rating.
  10. ## MANDATORY EXECUTION RULES (READ FIRST):
  11. ### Universal Rules:
  12. - 🛑 NEVER generate content without user input
  13. - 📖 CRITICAL: Read the complete step file before taking any action
  14. - 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
  15. - 📋 YOU ARE A FACILITATOR, not a content generator
  16. - ✅ YOU MUST ALWAYS SPEAK OUTPUT In your Agent communication style with the config `{communication_language}`
  17. - ✅ YOU MUST ALWAYS WRITE all artifact and document content in `{document_output_language}`
  18. ### Role Reinforcement:
  19. - ✅ You are a Validation Architect and Quality Assurance Specialist
  20. - ✅ If you already have been given communication or persona patterns, continue to use those while playing this new role
  21. - ✅ We engage in systematic validation, not collaborative dialogue
  22. - ✅ You bring analytical rigor and document quality expertise
  23. - ✅ This step runs autonomously - no user input needed
  24. - ✅ Uses Advanced Elicitation for multi-perspective evaluation
  25. ### Step-Specific Rules:
  26. - 🎯 Focus ONLY on holistic document quality assessment
  27. - 🚫 FORBIDDEN to validate individual components (done in previous steps)
  28. - 💬 Approach: Multi-perspective evaluation using Advanced Elicitation
  29. - 🚪 This is a validation sequence step - auto-proceeds when complete
  30. ## EXECUTION PROTOCOLS:
  31. - 🎯 Use Advanced Elicitation for multi-perspective assessment
  32. - 🎯 Evaluate document flow, dual audience, BMAD principles
  33. - 💾 Append comprehensive assessment to validation report
  34. - 📖 Display "Proceeding to next check..." and load next step
  35. - 🚫 FORBIDDEN to pause or request user input
  36. ## CONTEXT BOUNDARIES:
  37. - Available context: Complete PRD file, validation report with findings from steps 1-10
  38. - Focus: Holistic quality - the WHOLE document
  39. - Limits: Don't re-validate individual components, don't pause for user input
  40. - Dependencies: Steps 1-10 completed - all systematic checks done
  41. ## MANDATORY SEQUENCE
  42. **CRITICAL:** Follow this sequence exactly. Do not skip, reorder, or improvise unless user explicitly requests a change.
  43. ### 1. Attempt Sub-Process with Advanced Elicitation
  44. **Try to use Task tool to spawn a subprocess using Advanced Elicitation:**
  45. "Perform holistic quality assessment on this PRD using multi-perspective evaluation:
  46. **Advanced Elicitation workflow:**
  47. Invoke the `bmad-advanced-elicitation` skill
  48. **Evaluate the PRD from these perspectives:**
  49. **1. Document Flow & Coherence:**
  50. - Read entire PRD
  51. - Evaluate narrative flow - does it tell a cohesive story?
  52. - Check transitions between sections
  53. - Assess consistency - is it coherent throughout?
  54. - Evaluate readability - is it clear and well-organized?
  55. **2. Dual Audience Effectiveness:**
  56. **For Humans:**
  57. - Executive-friendly: Can executives understand vision and goals quickly?
  58. - Developer clarity: Do developers have clear requirements to build from?
  59. - Designer clarity: Do designers understand user needs and flows?
  60. - Stakeholder decision-making: Can stakeholders make informed decisions?
  61. **For LLMs:**
  62. - Machine-readable structure: Is the PRD structured for LLM consumption?
  63. - UX readiness: Can an LLM generate UX designs from this?
  64. - Architecture readiness: Can an LLM generate architecture from this?
  65. - Epic/Story readiness: Can an LLM break down into epics and stories?
  66. **3. BMAD PRD Principles Compliance:**
  67. - Information density: Every sentence carries weight?
  68. - Measurability: Requirements testable?
  69. - Traceability: Requirements trace to sources?
  70. - Domain awareness: Domain-specific considerations included?
  71. - Zero anti-patterns: No filler or wordiness?
  72. - Dual audience: Works for both humans and LLMs?
  73. - Markdown format: Proper structure and formatting?
  74. **4. Overall Quality Rating:**
  75. Rate the PRD on 5-point scale:
  76. - Excellent (5/5): Exemplary, ready for production use
  77. - Good (4/5): Strong with minor improvements needed
  78. - Adequate (3/5): Acceptable but needs refinement
  79. - Needs Work (2/5): Significant gaps or issues
  80. - Problematic (1/5): Major flaws, needs substantial revision
  81. **5. Top 3 Improvements:**
  82. Identify the 3 most impactful improvements to make this a great PRD
  83. Return comprehensive assessment with all perspectives, rating, and top 3 improvements."
  84. **Graceful degradation (if no Task tool or Advanced Elicitation unavailable):**
  85. - Perform holistic assessment directly in current context
  86. - Read complete PRD
  87. - Evaluate document flow, coherence, transitions
  88. - Assess dual audience effectiveness
  89. - Check BMAD principles compliance
  90. - Assign overall quality rating
  91. - Identify top 3 improvements
  92. ### 2. Synthesize Assessment
  93. **Compile findings from multi-perspective evaluation:**
  94. **Document Flow & Coherence:**
  95. - Overall assessment: [Excellent/Good/Adequate/Needs Work/Problematic]
  96. - Key strengths: [list]
  97. - Key weaknesses: [list]
  98. **Dual Audience Effectiveness:**
  99. - For Humans: [assessment]
  100. - For LLMs: [assessment]
  101. - Overall dual audience score: [1-5]
  102. **BMAD Principles Compliance:**
  103. - Principles met: [count]/7
  104. - Principles with issues: [list]
  105. **Overall Quality Rating:** [1-5 with label]
  106. **Top 3 Improvements:**
  107. 1. [Improvement 1]
  108. 2. [Improvement 2]
  109. 3. [Improvement 3]
  110. ### 3. Report Holistic Quality Findings to Validation Report
  111. Append to validation report:
  112. ```markdown
  113. ## Holistic Quality Assessment
  114. ### Document Flow & Coherence
  115. **Assessment:** [Excellent/Good/Adequate/Needs Work/Problematic]
  116. **Strengths:**
  117. {List key strengths}
  118. **Areas for Improvement:**
  119. {List key weaknesses}
  120. ### Dual Audience Effectiveness
  121. **For Humans:**
  122. - Executive-friendly: [assessment]
  123. - Developer clarity: [assessment]
  124. - Designer clarity: [assessment]
  125. - Stakeholder decision-making: [assessment]
  126. **For LLMs:**
  127. - Machine-readable structure: [assessment]
  128. - UX readiness: [assessment]
  129. - Architecture readiness: [assessment]
  130. - Epic/Story readiness: [assessment]
  131. **Dual Audience Score:** {score}/5
  132. ### BMAD PRD Principles Compliance
  133. | Principle | Status | Notes |
  134. |-----------|--------|-------|
  135. | Information Density | [Met/Partial/Not Met] | {notes} |
  136. | Measurability | [Met/Partial/Not Met] | {notes} |
  137. | Traceability | [Met/Partial/Not Met] | {notes} |
  138. | Domain Awareness | [Met/Partial/Not Met] | {notes} |
  139. | Zero Anti-Patterns | [Met/Partial/Not Met] | {notes} |
  140. | Dual Audience | [Met/Partial/Not Met] | {notes} |
  141. | Markdown Format | [Met/Partial/Not Met] | {notes} |
  142. **Principles Met:** {count}/7
  143. ### Overall Quality Rating
  144. **Rating:** {rating}/5 - {label}
  145. **Scale:**
  146. - 5/5 - Excellent: Exemplary, ready for production use
  147. - 4/5 - Good: Strong with minor improvements needed
  148. - 3/5 - Adequate: Acceptable but needs refinement
  149. - 2/5 - Needs Work: Significant gaps or issues
  150. - 1/5 - Problematic: Major flaws, needs substantial revision
  151. ### Top 3 Improvements
  152. 1. **{Improvement 1}**
  153. {Brief explanation of why and how}
  154. 2. **{Improvement 2}**
  155. {Brief explanation of why and how}
  156. 3. **{Improvement 3}**
  157. {Brief explanation of why and how}
  158. ### Summary
  159. **This PRD is:** {one-sentence overall assessment}
  160. **To make it great:** Focus on the top 3 improvements above.
  161. ```
  162. ### 4. Display Progress and Auto-Proceed
  163. Display: "**Holistic Quality Assessment Complete**
  164. Overall Rating: {rating}/5 - {label}
  165. **Proceeding to final validation checks...**"
  166. Without delay, read fully and follow: {nextStepFile} (step-v-12-completeness-validation.md)
  167. ---
  168. ## 🚨 SYSTEM SUCCESS/FAILURE METRICS
  169. ### ✅ SUCCESS:
  170. - Advanced Elicitation used for multi-perspective evaluation (or graceful degradation)
  171. - Document flow & coherence assessed
  172. - Dual audience effectiveness evaluated (humans and LLMs)
  173. - BMAD PRD principles compliance checked
  174. - Overall quality rating assigned (1-5 scale)
  175. - Top 3 improvements identified
  176. - Comprehensive assessment reported to validation report
  177. - Auto-proceeds to next validation step
  178. - Subprocess attempted with graceful degradation
  179. ### ❌ SYSTEM FAILURE:
  180. - Not using Advanced Elicitation for multi-perspective evaluation
  181. - Missing document flow assessment
  182. - Missing dual audience evaluation
  183. - Not checking all BMAD principles
  184. - Not assigning overall quality rating
  185. - Missing top 3 improvements
  186. - Not reporting comprehensive assessment to validation report
  187. - Not auto-proceeding
  188. **Master Rule:** This evaluates the WHOLE document, not just components. Answers "Is this a good PRD?" and "What would make it great?"