The Agent Skills Directory

[SAFE]: The skill defines a set of research-oriented commands for financial analysis. The instructions are transparent and follow a logical workflow for data retrieval and analysis without requesting sensitive permissions or performing hidden actions.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external news sources (EastMoney, Sina), creating a potential surface for indirect prompt injection.
Ingestion points: Untrusted content enters the agent context via the get_market_news function referenced in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between its primary instructions and content retrieved from news sources.
Capability inventory: No high-risk capabilities such as file system writes, network exfiltration (beyond the defined tool calls), or arbitrary command execution were identified in the skill scripts.
Sanitization: The skill lacks explicit sanitization or validation protocols for the external news data before it is processed by the agent.

stock