multi-platform-apps-multi-platform
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied feature specifications through the
$ARGUMENTSvariable and interpolates them directly into instruction sets for multiple sub-agents (e.g.,backend-architect,ui-ux-designer). Malicious instructions embedded in the input could potentially influence the sub-agents' design and code generation outputs. - Ingestion points: User input is received via the
$ARGUMENTSplaceholder inSKILL.mdand passed to sub-agents responsible for architecture, design, and implementation. - Boundary markers: The instructions lack explicit delimiters (e.g., XML tags or triple quotes) or protective instructions to ensure the agent ignores potential commands within the user-provided text.
- Capability inventory: The skill utilizes a
Tasktool to delegate complex software engineering tasks, including API design and cross-platform implementation (Web, iOS, Android, Desktop), which provides a high-impact surface for successful injections. - Sanitization: No sanitization, validation, or filtering of the user-provided input is performed before it is used in prompt interpolation.
Audit Metadata