The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions and does not contain any executable scripts or binary files. It references a project-specific context file (.claude/product-marketing-context.md) for grounding, which is standard behavior for an agent skill and poses no security risk.
[PROMPT_INJECTION]: The skill instructs the agent to analyze potentially untrusted data from web search results, call transcripts, and survey data. This creates a surface for indirect prompt injection, which is inherent to the use case of marketing research. Ingestion points: External transcripts, survey data, and web search results. Boundary markers: Absent. Capability inventory: Natural language processing and web search; no code execution or sensitive file access. Sanitization: Absent.

content-strategy