marketing-ideas
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as a static library of marketing information. It does not contain executable scripts, network requests, or sensitive data access beyond its intended purpose.
- [DATA_EXPOSURE]: The skill references a local file path
.claude/product-marketing-context.mdto retrieve product-specific context. This is a common pattern for maintaining context within a workspace and does not involve exfiltration of data to external servers. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface via the
.claude/product-marketing-context.mdfile and user input. - Ingestion points:
.claude/product-marketing-context.mdand runtime user queries. - Boundary markers: None explicitly defined for the file read.
- Capability inventory: The skill is limited to generating text recommendations; it lacks capabilities for subprocess execution, file writing, or network operations.
- Sanitization: Not present, but the lack of high-privilege capabilities mitigates the risk of exploitation through injected instructions in the context file.
Audit Metadata