sirv-ai-studio
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides configuration instructions to download the
sirv-ai-mcp-serverpackage from the NPM registry during setup. - [REMOTE_CODE_EXECUTION]: The recommended configuration for the MCP server uses
npx -y sirv-ai-mcp-server, which downloads and executes remote code on the host machine. - [COMMAND_EXECUTION]: The skill instructs the user to configure the agent's environment to execute shell-level commands (via
npx) to facilitate the MCP tool integration. - [DYNAMIC_EXECUTION]: The use of
npxinvolves the runtime execution of a package that is not pre-installed, representing dynamic code execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes external content through its image and video manipulation tools, creating a vulnerability surface where malicious data could influence agent behavior.
- Ingestion points: Multiple tools in
references/mcp-tools.mdandreferences/asset-product-tools.mdaccept animage_urlparameter from external sources. - Boundary markers: The instructions lack specific delimiters or warnings for the agent to ignore potentially malicious instructions embedded in image metadata or descriptors.
- Capability inventory: The skill tools perform complex image transformations (upscaling, lifestyle generation, etc.) and return results to the conversation.
- Sanitization: There is no evidence of sanitization or safety checks applied to the content retrieved from external URLs.
Audit Metadata