email-design
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references an installation script at
https://cli.inference.sh. This URL belongs to the skill author's infrastructure and is used to distribute the necessaryinfshCLI tool. - [REMOTE_CODE_EXECUTION]: The Quick Start section includes a command to pipe a remote script from
https://cli.inference.shdirectly into the shell (sh). This is the documented installation method for the vendor's utility. - [COMMAND_EXECUTION]: The skill utilizes the
infshcommand-line interface to execute image generation tasks andnpxto manage skill dependencies. - [PROMPT_INJECTION]: The skill uses tools that process user-provided strings and HTML for image generation, creating a surface for indirect prompt injection.
- Ingestion points: Markdown examples show interpolation of HTML and text prompts into
infsh app runJSON inputs. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the templates.
- Capability inventory: The skill is granted permission to run the
infshCLI via theBashtool. - Sanitization: No sanitization or validation of the input strings is performed within the provided skill definition.
Audit Metadata