mage-face-swap
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes official vendor tools and infrastructure (mage.space) to provide its service. All tool calls and external references point to the vendor's own domain, representing standard functionality.
- [SAFE]: The instructions include explicit safety and ethical guidelines prohibiting the creation of harmful deepfakes, requiring user permission for likeness use, and forbidding harassment or deception.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided image URLs and local files through the
create_uploadtool. This represents an ingestion surface for untrusted data, but the risk is mitigated because the input is used as structured arguments for tool execution rather than being directly interpreted as agent instructions. Boundary markers are established via the JSON schema structure in the tool calls. - [COMMAND_EXECUTION]: The skill mentions running commands to measure files or upload local files. These operations are conducted through defined agent tools (
create_upload) rather than arbitrary shell execution, maintaining a controlled environment.
Audit Metadata