mental-wellness
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input regarding moods, life situations, and emotional states to generate wellness templates and journaling prompts.
- Ingestion points: User input provided as arguments or responses to menu options in SKILL.md and templates/prompt-main.md.
- Boundary markers: The instructions lack explicit structural delimiters to isolate user-supplied strings from the internal instruction set, relying instead on natural language parsing.
- Capability inventory: The skill's primary capabilities are text generation and writing markdown files as documented in the Output Format section of SKILL.md.
- Sanitization: The skill features a detailed 'Safety Screening' protocol (SKILL.md) that monitors user input for specific 'red flag' keywords related to self-harm or medical emergencies, triggering immediate cessation and emergency resource referral.
Audit Metadata