sms
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to proactively read external context files such as
.agents/product-marketing.md,.claude/product-marketing.md, orproduct-marketing-context.mdat the start of a task. This constitutes an ingestion point for untrusted data that may be controlled by an attacker or contained within a repository the agent is invited to analyze. - Ingestion points: Identified in
SKILL.mdunder the 'Before Starting' section. - Boundary markers: The instructions lack explicit delimiting or warnings to ignore embedded instructions within these ingested files.
- Capability inventory: The skill does not define direct shell or code execution tools, but it generates marketing strategies and copy that influence user-facing communications.
- Sanitization: No validation or sanitization of the content from these external files is specified before the agent processes it into its context.
Audit Metadata