/generate-prp

description: Generate comprehensive PRP (Product Requirements & Plans) with thorough research and validation
argument-hint: [feature description or user story]
allowed-tools: TodoWrite, Read, Write, Glob, Grep, Bash, Task, WebSearch, WebFetch

Create PRP

Feature file: $ARGUMENTS

Generate PRP (Product Requirements & Plans) through validated research and codebase analysis.

Workflow Summary

Two-phase approach: validate completeness (Phase 1), then research (Phase 2).

**CRITICAL**: PRP must contain ALL context - research findings, documentation URLs, code examples. The executor agent sees only the final PRP document, not your research process.

Research Process

**Phase 1: Initial Discovery & Task Validation** (Validate task completeness before deep research)

1. **Preflight Analysis** (Use subagent: `bp:preflight-prp`)

• Quick scan of project structure for similar features/patterns
• Analyze user's task description for business logic completeness
• Identify gaps in user requirements and missing business logic details:
• User flows and interaction patterns
• Data requirements and relationships
• Integration points with existing features
• Edge cases and error scenarios
• UI/UX expectations and constraints
• **Language Guidelines for Questions**:
• Ask clarification questions in the same language the user wrote the initial task
• Wait for user responses and analyze thoroughly
• Generate targeted clarification questions if gaps identified
• Make proceed/clarify recommendation with clear reasoning

2. **Decision Gate**:

• **IF** bp:preflight-prp recommends PROCEED → Continue to Phase 2
• **IF** bp:preflight-prp identifies gaps → Stop and ask clarifying

questions

• **ONLY** continue to comprehensive research after user provides missing

details

• Use surface discovery findings to inform Phase 2 research focus

**Phase 2: Comprehensive Research Phase** (After task validation - Codebase first, then smart external research)

1. **Codebase Analysis** (Use subagent: `bp:codebase-research`)

• Search for similar features/patterns in the codebase
• Identify files to reference in PRP
• Note existing conventions to follow
• Check test patterns for validation approach
• **CRITICAL**: Document what components/libraries/patterns already exist
• **ASSESS**: Determine knowledge gaps that truly need external research

2. **Smart External Research Decision** (Evaluate AFTER codebase analysis)

**FIRST: Analyze codebase findings to determine if external research is needed:**

**SKIP External Research if:**

• ✅ Similar components/patterns found in codebase (internal project

components)

• ✅ Clear implementation path from existing code
• ✅ Standard CRUD/UI operations using existing patterns
• ✅ Internal utility functions/services already available

**PROCEED with External Research ONLY if:**

• ❌ New external npm/library integration needed (get current docs)
• ❌ Existing external library usage but complex/undocumented features

needed

• ❌ Complex algorithm or pattern not in codebase
• ❌ Security/performance considerations beyond current code
• ❌ External API integration without existing examples
• ❌ **No similar patterns/components found in codebase** (need external

examples)

**If External Research is needed** (Use subagent: `bp:research-agent`):

• Focus ONLY on missing knowledge gaps identified above
• External npm/library documentation for NEW packages or complex features
• Best practices for COMPLEX patterns not in codebase
• Security considerations for NEW external integrations
• **AVOID**: Researching internal project components (use codebase instead)
• **Agent returns all findings directly in response context**

3. **Technical Clarification** (Use if needed after research completion)

• **ONLY** for technical implementation details, not business logic
• Specific patterns to mirror and where to find them?
• Integration requirements and where to find them?
• Which existing service to use and its file path?
• Confirm if external research is truly needed for identified gaps

Language Guidelines

User Interaction Language

• **Questions & Communication**: Ask all clarification questions in the same language the user wrote the initial task
• **Analysis & Discussion**: Continue using the user's language throughout the discovery and research phases

PRP Document Language

• **Final Document**: Always write the PRP document in English for consistency and international team compatibility
• **User Response Translation**: When incorporating user responses into the PRP, translate them to English while preserving the original meaning
• **Code Examples**: Always use English comments and variable names in technical examples
• **Technical Terms**: Use standard English technical terminology in the final document

PRP Generation

**IMPORTANT**: Read and follow the template structure from `docs/templates/prp_document_template.md` exactly.

If template doesn't exist, tell user to run `/bp:init` first to install templates.

**CRITICAL BEFORE WRITING**: 1. Complete all research phases (Phase 1 and Phase 2) 2. Gather ALL context from codebase analysis 3. Document validation commands from project config files 4. THEN write the PRP following the template structure

Output

Save as: `docs/prps/{feature-name}.md`

Quality Checklist

• [ ] All necessary context included
• [ ] Validation gates are executable by AI
• [ ] References existing patterns
• [ ] Clear implementation path
• [ ] Error handling documented

Score the PRP on a scale of 1-10 (confidence level to succeed in one-pass implementation using claude codes)

Task Breakdown Generation

**Final Step: Generate Implementation Tasks**

After completing the PRP document, automatically generate a detailed task breakdown:

1. **Task Decomposition** (Use subag