churn-prevention
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill contains documentation and instructional content related to business operations such as churn reduction and payment recovery. It references well-known industry services like Stripe and PostHog for implementation guidance.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to ingest content from external files for product context, creating an ingestion surface for untrusted data. However, the skill does not define any executable capabilities that could be leveraged for an exploit, making this a standard operational pattern for this type of consultative agent. Ingestion points: The skill reads from .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md as specified in SKILL.md. Boundary markers: No specific delimiters or safety instructions are mentioned for the external content. Capability inventory: The skill definition includes no subprocess calls, network operations, or file-writing capabilities across its files. Sanitization: No sanitization or filtering logic is specified for the ingested content.
Audit Metadata