Skip to content

Iterative Design Loop

Three-step process for handling new information that impacts the design: scan, analyze, route.

The product design workflow supports iteration at any point. When new information arrives -- usability test results, stakeholder feedback, technical constraints, or market intelligence -- these three commands detect, assess, and route you back to the right phase.

The loop

New Input --> /pd-inbox-scan --> /pd-analyze-changes --> /pd-route-decision
                                                              |
                                            Return to pd-1, pd-2, pd-3, etc.

Step 1: Inbox Scan

Command: /pd-inbox-scan

Scans the 00-Inbox/ folder for files newer than the last scan. Each item is classified by type and urgency:

Type Urgency Example
Usability test results HIGH usability-test-results.pdf
Stakeholder feedback MEDIUM stakeholder-review-notes.md
Technical constraint HIGH api-feasibility-report.md
Scope change request HIGH change-request-new-feature.md
Market intelligence LOW competitor-analysis.md
User research MEDIUM interview-transcript-05.pdf

Updates .design-state.yaml with pending items and their urgency levels. HIGH urgency items require immediate analysis before continuing current phase work.


Step 2: Analyze Changes

Command: /pd-analyze-changes

Reads each pending item, extracts findings, and applies the impact matrix to determine which artifacts are affected:

Change type PRD Prototype SPECchen Return to
Usability test (critical) -- HIGH -- pd-3
Stakeholder feedback (scope) HIGH HIGH HIGH pd-2
User research (invalidating) HIGH HIGH HIGH pd-1
Tech constraint (breaking) HIGH HIGH HIGH pd-2
Bug report (prototype) -- HIGH -- pd-3

When multiple items have different recommendations, the earliest affected phase wins (pd-1 < pd-2 < pd-3 < pd-4).

Produces an impact report with aggregated recommendations and multiple routing options.


Step 3: Route Decision

Command: /pd-route-decision [--option N | --continue]

Presents the routing options from the analysis and captures your decision:

  1. Return to recommended phase -- Full or targeted rework
  2. Return to alternative phase -- Address subset of issues
  3. Split into parallel tracks -- Quick fixes vs. larger changes
  4. Continue current phase -- Defer items to backlog
  5. Manual override -- Custom routing

Upon selection: - Increments the iteration counter in .design-state.yaml - Resets quality gates for the target phase and all subsequent phases - Bumps artifact versions for affected deliverables - Logs the decision in iteration history - Launches the target phase skill automatically


When to check for iterations

The design state should be checked: - Before starting each phase (built into every /pd-* skill) - When usability tests reveal issues - When stakeholder feedback arrives - When technical constraints are discovered

The .design-state.yaml file tracks the full iteration history, artifact hashes, quality gate status, and all routing decisions for process learning.