The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a potential attack surface by instructing the agent to ingest data from external context files, though its lack of dangerous capabilities makes this a low-risk pattern.
Ingestion points: The skill instructs the agent to read .agents/product-marketing-context.md or .claude/product-marketing-context.md to inform its responses (found in SKILL.md).
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within these context files.
Capability inventory: The skill does not contain scripts or instructions for file system modifications, network requests, or shell command execution; it only generates JSON-LD code blocks for user implementation.
Sanitization: There is no mention of sanitizing or escaping the data retrieved from the context files before it is processed by the agent.
[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were identified. All external references point to well-known search engine validation tools and schema.org documentation.

schema-markup