kinetic-action-visual-master
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and reference materials are designed for creative text generation. It does not utilize any tools, perform network requests, or execute scripts. All content is localized to the agent's prompt context.
- [INDIRECT_PROMPT_INJECTION]: The skill contains a functional surface for ingesting user-provided "named anchors" into generated prompt blocks. Although the risk is negligible due to the lack of exploitable tools, the surface is defined by the following characteristics:
- Ingestion points: User-provided "named anchors" (SKILL.md, Step 2).
- Boundary markers: The skill instructs the agent to preserve user terms verbatim, lacking explicit delimiters to isolate untrusted input.
- Capability inventory: No file system, network, or command execution capabilities are available to this skill.
- Sanitization: No specific filtering or escaping is applied to user-supplied anchors.
Audit Metadata