marketing-ideas
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were detected. The skill consists of instructional text and marketing reference data.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read a specific local file for context, creating a potential ingestion point for untrusted instructions. * Ingestion points: Reads product context from .claude/product-marketing-context.md. * Boundary markers: Absent. * Capability inventory: None. The skill has no access to the shell, network, or file system for writing. * Sanitization: Absent. This finding is rated SAFE as there are no exploitable capabilities present in the skill.
Audit Metadata