The Agent Skills Directory

[SAFE]: The skill's primary function is to generate JSON-LD schema markup based on user input and project context. It does not perform any dangerous operations such as command execution, network requests to untrusted domains, or sensitive file access.
[INDIRECT_PROMPT_INJECTION]: The skill includes instructions to read .claude/product-marketing-context.md to gather project information. This represents a potential surface for indirect prompt injection if the context file contains malicious instructions. However, the risk is negligible as the skill lacks high-privilege capabilities (like file writes or network exfiltration) to exploit such an injection.
Ingestion points: .claude/product-marketing-context.md (referenced in SKILL.md and its translations).
Boundary markers: Absent; the agent is simply told to "read it before asking questions."
Capability inventory: Limited to text and JSON generation. No subprocess execution, file system modification, or network tool access listed.
Sanitization: None specified for the content read from the context file.
[SAFE]: External links provided (Google Rich Results Test, Schema.org Validator) are well-known, trusted industry tools for validating structured data.

schema-markup