churn-prevention
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests external business context files and user-supplied metrics without specifying safety boundaries, creating a surface for indirect prompt injection.
- Ingestion points: Context is gathered from .agents/product-marketing.md, .claude/product-marketing.md, and legacy product-marketing-context.md files.
- Boundary markers: The instructions lack guidance for the agent to use delimiters or 'ignore embedded instructions' warnings when processing this external content.
- Capability inventory: The agent is guided to use tools such as stripe, customer-io, posthog, mixpanel, ga4, and segment for implementation.
- Sanitization: There is no requirement for the agent to sanitize or validate the external content before using it to drive strategies.
Audit Metadata