Helps you recognize validation loops, one-sided framing, and pseudo-diagnoses in AI-generated relationship advice.
AI Relationship Safety Guide
Evaluate AI relationship advice critically
Best for
- ▸Recognize validation loops and confirmation bias in AI relationship advice
- ▸Identify pseudo-diagnoses and partner labeling in AI output
- ▸Understand when AI replaces rather than supports human connection
- ▸Evaluate AI-drafted relationship messages for appropriateness
What you'll get
- ▸Validation loop detection showing AI elaborating accusatory framing without questioning assumptions
- ▸Pseudo-diagnosis flag for AI labeling a partner as "narcissistic" from a one-sided 3-paragraph description
- ▸Safe zone assessment confirming AI-drafted apology message as helpful de-escalation support
- ▸One-sidedness report recommending couples therapy after AI gave a verdict on a trust breach situation
AI-generated relationship advice, couples communication question, or concern about AI partnership guidance
Risk assessment with one-sidedness check, validation loop detection, pseudo-diagnosis flags, and therapist referral guidance
What's inside
“You are a Relationship Safety Guide. You help users evaluate AI-generated relationship advice critically by identifying where it becomes harmful through one-sided perspective, pseudo-diagnosis, and therapy replacement. - You flag structural limitations (AI only hears one side; cannot infer partner i...”
Covers
Not designed for ↓
- ×Providing relationship advice or couples counseling
- ×Replacing licensed therapists or counselors
- ×Diagnosing partners with personality disorders
- ×Making separation or divorce decisions
SupaScore
89.4▼
Evidence Policy
Standard: no explicit evidence policy.
Research Foundation: 8 sources (7 paper, 1 industry frameworks)
This skill was developed through independent research and synthesis. SupaSkills is not affiliated with or endorsed by any cited author or organisation.
Version History
v5.5 final distill
Works well with
© 2026 Kill The Dragon GmbH. This skill and its system prompt are protected by copyright. Unauthorised redistribution is prohibited. Terms of Service · Legal Notice