cross-platform-campaign
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of user-provided campaign briefs to generate social media copy, which is then passed to drafting tools.
- Ingestion points: User-supplied campaign briefs and marketing objectives (processed in the first step of the workflow in SKILL.md).
- Boundary markers: The instructions do not define clear delimiters (e.g., XML tags or backticks) or include explicit instructions for the model to ignore potential commands embedded within the user's brief.
- Capability inventory: The skill uses
social_getSocialMediaAccountsto read account info andsocial_createSocialMediaPostto create drafts across connected platforms. - Sanitization: There is no logic specified to sanitize or escape user-provided text before it is interpolated into post content.
- [EXTERNAL_DOWNLOADS]: The skill relies on an external Model Context Protocol (MCP) server for its core functionality.
- Evidence: The configuration in
agents/openai.yamlspecifies a dependency onhttps://apikit.simplified.com/mcpto provide the image and social media tools. This is a reference to an official endpoint for the Simplified marketing service.
Audit Metadata