vega
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- PROMPT_INJECTION (SAFE): No override markers, role-play instructions, or attempts to bypass safety filters were detected.
- DATA_EXFILTRATION (SAFE): There are no commands for accessing sensitive local files or sending data to external servers.
- REMOTE_CODE_EXECUTION (SAFE): The skill does not download or execute remote scripts; the schema URL is a standard reference for visualization validation.
- INDIRECT_PROMPT_INJECTION (SAFE): The skill has a data ingestion surface (numeric data arrays in SKILL.md) but lacks exploitable capabilities like subprocess execution or file writing. It uses markdown code blocks as boundary markers to contain output specifications.
- PRIVILEGE_ESCALATION (SAFE): No commands for elevating user permissions or modifying system files were found.
Audit Metadata