Workflows — chain skills and subagents into a single run

Workflows chain multiple skills and subagents into one run. Drop your inputs, run the workflow, and get back a structured deliverable — without stitching outputs together yourself. Personal PM OS comes with 7 workflows out of the box, covering discovery, strategy, and planning.

When to use this

  • You have a multi-step PM task that combines several skills end-to-end.
  • You want parallel review or research from multiple subagents in a single run.
  • You need raw inputs (a problem statement, scattered notes, competitor URLs) turned into a structured deliverable.

Running a workflow

  1. Drop your input files into the workflow's input folder.
  2. Invoke the workflow in Claude Code.
  3. Review the structured outputs the workflow writes back.

How it works

A workflow is a defined sequence that calls skills and subagents in order — sometimes in parallel — and threads their outputs together. Problem-to-PRD takes a raw problem statement and runs it through framing, user-impact analysis, and drafting skills to land a shipped PRD in one pass. Competitive Intel Pack runs three subagents in parallel — each researching a different angle — then merges their findings into a single competitor profile.

You can customize any of the included workflows, or build your own to match how your team actually works. You invoke a workflow the same way you invoke a skill — the difference is that one run can span multiple steps without you stitching outputs together.

Workflows included with Personal PM OS

WorkflowWhat it doesTime saved
Batch Interview Analysis5–10 interview transcripts → snapshots, synthesis, personas, opportunity tree7–10 hrs → 15–20 min
Problem to PRDProblem statement + evidence → JTBD, opportunity tree, PRD, user stories4–6 hrs → 30–45 min
Multi-Perspective ReviewA doc → 7 stakeholder reviews + consensus + tensions + blockers2+ hrs → 15 min
Competitive Intel Pack3–5 competitors → executive summary, profiles, feature matrix, SWOT4 hrs → 30 min
Hypothesis Tester"Conversion dropped 20%" → 5 hypotheses investigated + likely root cause2 hrs → 15 min
Market Sizing AnalyzerCategory + geography → TAM/SAM/SOM from 5 methods, triangulated4 hrs → 45 min
Voice of Customer Analysis2+ VoC sources → themes, prioritized pains, opportunities3 hrs → 30 min