The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external markdown files on GitHub (github.com/coreyhaines31/marketingskills) for tool registries and integration guides. These links are informational and point to a well-known service.
[INDIRECT_PROMPT_INJECTION]: The skill includes instructions to read and incorporate context from local files, which represents an indirect prompt injection attack surface if those files contain untrusted content. Ingestion points: Reads project-specific data from .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md. Boundary markers: Absent; the skill does not specify delimiters to separate untrusted context from instructions. Capability inventory: Mentions the use of powerful CLI tools including stripe, customer-io, posthog, and segment. Sanitization: Absent; there is no instruction to validate or filter the content of the marketing context files.

churn-prevention