submit-anime-project
Pass
Audited by Gen Agent Trust Hub on Feb 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- Data Exposure & Exfiltration (LOW): The skill reads local files and sends them to a network endpoint. Although the default is localhost, the 'api_base' parameter allows redirection to any external URL.
- Indirect Prompt Injection (LOW): The skill processes untrusted input from 'seedance_project_tasks.json' which controls file access. Ingestion points: 'seedance_project_tasks.json' in the project directory. Boundary markers: None. Capability inventory: Local file read (via 'open') and network POST (via '/api/tasks/push'). Sanitization: The skill lacks path validation when joining directory paths with task-provided relative paths, allowing potential path traversal.
Audit Metadata