workflow-designer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- Prompt Injection (SAFE): Instructions are strictly focused on engineering task decomposition and do not contain bypasses, role-play injections, or instructions to ignore system safety guidelines.
- Data Exposure & Exfiltration (SAFE): No sensitive file paths, credentials, or network exfiltration patterns were detected; the skill only generates text checklists.
- Remote Code Execution (SAFE): The skill consists entirely of markdown instructions and does not include any scripts or executable content.
- Indirect Prompt Injection (SAFE): Ingestion point: User task description in SKILL.md; Boundary markers: Absent; Capability inventory: No file-system, network, or shell access (text generation only); Sanitization: Absent. No malicious capabilities exist to be exploited via this surface.
Audit Metadata