The Agent Skills Directory

[DATA_EXFILTRATION]: The skill attempts to read local files such as .agents/product-marketing-context.md or .claude/product-marketing-context.md. These files are intended to provide context for marketing tasks and do not target sensitive user credentials or system configurations.
[EXTERNAL_DOWNLOADS]: The skill uses WebFetch to retrieve information from the Apple App Store (apps.apple.com) and Google Play Store (play.google.com). These are well-known services, and the data retrieved is limited to public application metadata required for auditing purposes.
[PROMPT_INJECTION]: The skill processes untrusted metadata from app store listings which presents a surface for indirect prompt injection.
Ingestion points: Data fetched from apps.apple.com and play.google.com via WebFetch (SKILL.md)
Boundary markers: Absent
Capability inventory: Screenshot capture, web fetching, and local file reading (SKILL.md)
Sanitization: Absent
Despite the absence of formal boundaries or sanitization, the risk is minimized by the specific analytical scope of the audit and the use of reputable public data sources.

aso-audit