# Responsible Design Review

> Review product or AI work for human impact, consent, autonomy, fairness, dark patterns, vulnerable users, and data dignity.

- Skill: `mphaxise/responsible-design-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add mphaxise/responsible-design-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/mphaxise/responsible-design-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: mphaxise (https://skillmd.com/u/mphaxise)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/mphaxise/responsible-design-review

---


# Responsible Design Review

Use this skill when a product, feature, plan, workflow, or AI agent behavior needs a human-impact review before it ships or becomes a default.

This is a Praneet-specific hand-port enhancement. It builds on GStack design review and GBrain memory, with an added design leadership lens for responsible design and social ethics.

## Workflow

1. Define the human context:
   - who is affected
   - who has less power
   - what data or attention is being asked for
   - what could happen if the system is wrong, persuasive, or overused
2. Check responsible design dimensions:
   - consent and meaningful choice
   - user autonomy and agency
   - dark patterns or coercive defaults
   - vulnerable users and edge populations
   - bias, fairness, and representational harm
   - privacy, retention, deletion, and data dignity
   - explainability and recourse
   - escalation paths for sensitive outcomes
3. Review product incentives:
   - what the business wants
   - what the user wants
   - where those incentives diverge
4. Produce a decision:
   - proceed
   - proceed with mitigations
   - pause for research or policy review
   - do not ship in the current form
5. Save durable findings through `reports`, `capture`, or `outcome-memory` when the decision should affect future work.

## Guardrails

- Do not reduce responsible design to a checklist when real harm is plausible.
- Do not treat legal compliance as the same thing as ethical design.
- Do not bury serious concerns under polish recommendations.
- Be explicit when the review needs affected-user research.