skills/aipoch/medical-research-skills/bidirectional-multi-phenotype-mr-research-planner/Gen Agent Trust Hub
bidirectional-multi-phenotype-mr-research-planner
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided exposure and outcome families to generate structured research designs.
- Ingestion points: External user input defining disease or trait families is processed in Step 1 of SKILL.md.
- Boundary markers: The skill contains dedicated sections for 'Input Validation' and 'Out-of-scope' redirection to handle off-topic or inappropriate inputs.
- Capability inventory: The skill provides study planning logic and does not incorporate scripts or instructions for shell command execution, file system access, or network exfiltration.
- Sanitization: The skill uses logical boundaries and prompt-based filtering to manage inputs but lacks technical sanitization of the input strings.
Audit Metadata