signup
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions include a step to read a local context file (.agents/marketing-context.md) if it exists. While this represents a data ingestion surface where instructions could theoretically be embedded in the context file, the skill lacks any capabilities—such as shell access, network tools, or file-writing logic—that could be exploited.
- Ingestion points: SKILL.md (references .agents/marketing-context.md)
- Boundary markers: Absent
- Capability inventory: None; the skill is limited to design guidelines and textual design analysis.
- Sanitization: Absent
Audit Metadata