churn-prevention
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest local context from files such as
.agents/product-marketing.md, which creates an attack surface where malicious instructions embedded in these files could influence agent behavior. - Ingestion points: Specifically reads from
.agents/product-marketing.mdand.claude/product-marketing.md(as seen in SKILL.md). - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are provided for the ingested content.
- Capability inventory: The skill interacts with several external CLI tools and APIs, including
stripe,customer-io,posthog,mixpanel,ga4, andsegment. - Sanitization: There is no evidence of sanitization, validation, or filtering for the external context being processed.
- [EXTERNAL_DOWNLOADS]: The instructions direct the user to an external tool registry and implementation guides hosted on a personal GitHub repository (
github.com/coreyhaines31/marketingskills) which is not a verified or trusted vendor.
Audit Metadata