The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the spark command-line tool to perform actions such as listing emails, unsubscribing from mailing lists, and blocking senders. These operations are consistent with the skill's stated purpose of email management.
[DATA_EXPOSURE]: The skill involves reading email threads to review newsletter content. This data access is necessary for the audit functionality and is performed within the context of the email service provider's authorized CLI.
[PROMPT_INJECTION]: Potential surface for Indirect Prompt Injection exists because the skill processes untrusted email content via spark thread. However, the skill does not possess high-risk capabilities (like arbitrary network requests or local file system modifications) that could be exploited by malicious email content. Boundary markers and sanitization are handled by the underlying agent platform and the spark tool itself.

spark-recipe-newsletter-cleanup