Skip to content

PRD / RFE Workflow

The PRD / RFE workflow helps product managers create comprehensive Product Requirements Documents and systematically break them down into actionable Request for Enhancement (RFE) items. The goal is to produce requirements documentation that is detailed enough for a senior engineer or architect to accept for implementation.

  • You need to create a formal PRD for a new feature or product initiative.
  • You want to break a PRD into well-scoped RFEs that can be tracked as individual work items.
  • You want agent-assisted prioritization using frameworks like RICE, MoSCoW, or Value vs. Effort.
  • You want to submit RFEs directly to Jira from the workflow.

The workflow follows a structured progression from discovery through requirements, PRD creation, review loops, RFE breakdown, and submission:

prd.discover

prd.requirements

prd.create

prd.review

prd.create

rfe.breakdown

rfe.review

rfe.prioritize

rfe.submit

CommandPurpose
/prd.discoverProduct discovery — define the problem statement, research user pain points, analyze competitive landscape, document assumptions and success metrics.
/prd.requirementsTransform discovery insights into specific, testable requirements with user stories, acceptance criteria, and MoSCoW prioritization.
CommandPurpose
/prd.createDraft or update the PRD with executive summary, product vision, goals, success metrics, and detailed requirements.
/prd.reviewReview the PRD for completeness, technical feasibility, and architecture alignment.
/prd.sourcesList all data sources that informed the PRD (documents, code, research, external sources).
CommandPurpose
/rfe.breakdownDecompose the PRD into implementable RFEs with clear acceptance criteria, technical dependencies, and sizing estimates.
/rfe.reviewReview individual RFEs for technical feasibility, testability, and architecture alignment.
/rfe.prioritizePrioritize RFEs using RICE, MoSCoW, Value vs. Effort, and Kano frameworks. Generate an implementation roadmap.
/rfe.submitFormat RFEs and submit them to Jira (or provide manual submission instructions if Jira is not connected).
/rfe.speedrunAutomatically run the entire workflow from discovery to Jira submission, pausing only for critical questions.

The workflow engages specialized agents at each phase:

  • Bailey (Business Analyst) — Business analysis and requirements elicitation.
  • Morgan (Technical Writer) — Documentation quality, clarity, and structure.
  • Parker (Product Manager) — Market strategy, competitive analysis, RICE scoring, business value.
  • Quinn (Product Strategist) — Product strategy and roadmap planning.
  • Riley (Product Owner) — User stories, acceptance criteria, backlog management, MoSCoW prioritization.
  • Ryan (UX Researcher) — User insights grounded in available research studies.
  • Terry (Technical Writer) — Technical writing and documentation standards.

The discovery phase can pull from multiple sources when configured:

  • Google Drive — Access product documents, stakeholder notes, and related assets.
  • UXR MCP — Structured access to user research reports and findings.
  • User-uploaded files — Supporting materials uploaded directly to the session.
  • Code repositories — Access to relevant codebases for technical context.
  • Jira — When the Jira MCP integration is configured in your workspace, RFEs can be submitted directly as tickets. Without it, the workflow provides manual submission instructions.
ArtifactPath
Discovery documentartifacts/discovery.md
Requirements documentartifacts/requirements.md
Product Requirements Documentartifacts/prd.md
PRD quality checklistartifacts/prd-checklist.md
PRD review reportartifacts/prd-review-report.md
RFE master listartifacts/rfes.md
Individual RFE filesartifacts/rfe-tasks/RFE-001-*.md, RFE-002-*.md, etc.
Prioritization analysisartifacts/prioritization.md
Jira ticket mappingartifacts/jira-tickets.md

The workflow includes a built-in quality rubric that evaluates RFEs on five dimensions, each scored from 1 to 5:

CriterionWhat it measures
CompletenessStructural completeness and organization.
ProfessionalismProfessional perspective and strategic depth.
ToneLanguage quality and communicative tone.
PurposeClarity of purpose and stakeholder alignment.
ActionabilityActionability and testability of requirements.

After generating RFEs, the agent automatically evaluates them against this rubric and produces an aggregate score out of 25.

  • Use /rfe.speedrun for a quick pass. It runs the entire workflow end-to-end, only pausing for critical questions. Good for a first draft that you can refine.
  • Run review loops. The workflow supports iterating on both the PRD (/prd.review then /prd.create) and the RFEs (/rfe.review then /rfe.breakdown). Multiple passes improve quality.
  • Connect Jira early. If Jira MCP is configured in your workspace, /rfe.submit will create tickets automatically with proper field mapping and dependency links. Without it, you get manual submission instructions.
  • Leverage data sources. Connect Google Drive and UXR MCP during workspace setup to give the discovery phase access to existing research and documents.