explore-recipes
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install packages from the author's own repository (
signerlabs/shipswift-skills) and connect to the official vendor MCP server athttps://api.shipswift.app/mcp. - [PROMPT_INJECTION]: The skill is identified as having an indirect prompt injection surface as it retrieves content from an external API which is then presented within the agent's context.
- Ingestion points: Content returned by the
listRecipes,searchRecipes, andgetRecipetools. - Boundary markers: Absent; the instructions do not explicitly tell the agent to ignore or delimit instructions found within the fetched recipe code or descriptions.
- Capability inventory: The skill is used for browsing and retrieving code templates but does not contain internal scripts with dangerous capabilities like arbitrary command execution.
- Sanitization: Absent; no validation or sanitization process for the external API responses is defined in the skill documentation.
Audit Metadata