The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to aggregate and analyze untrusted user-generated content from the moomoo community feed, which creates a surface for indirect prompt injection attacks.
Ingestion points: Community posts and comments are retrieved via the https://ai-news-search.moomoo.com/stock_feed endpoint as described in the Workflow section.
Boundary markers: The skill instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent processes the retrieved text for sentiment classification.
Capability inventory: The skill is configured to use system tools like curl and openssl for network operations and data processing.
Sanitization: Although the skill mandates stripping HTML tags, it does not provide mechanisms to sanitize or escape malicious natural language instructions hidden within the post content.
[COMMAND_EXECUTION]: The metadata requires the presence of curl, openssl, and date binaries. These tools are used to perform HTTP requests to the vendor's API and to manipulate timestamps during data processing.

moomoo-comment-sentiment