The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze data from external sources, specifically App Store and Google Play listings, including user reviews and HTML content. This represents a potential surface for indirect prompt injection where malicious instructions could be hidden in store data. However, the skill explicitly includes defensive instructions for the agent to ignore any commands found within the processed data and treats this content as unreliable.
[EXTERNAL_DOWNLOADS]: The skill uses the WebFetch tool to retrieve information from well-known and trusted domains, specifically apps.apple.com and play.google.com. These operations are necessary for the primary function of the skill and do not involve downloading or executing untrusted scripts.
[COMMAND_EXECUTION]: The skill includes instructions to use web search and navigation tools (like taking screenshots of a page) to gather visual data. These actions are scoped to the intended ASO audit task and do not involve arbitrary command execution or privilege escalation.
[DATA_EXPOSURE]: The skill attempts to read local project files such as .agents/product-marketing.md to gain context for its audit. This is a standard practice for providing context-aware analysis and does not involve exfiltrating sensitive credentials or personal information to external servers.

aso