The Agent Skills Directory

[SAFE]: The skill defines a legitimate and useful workflow for financial market analysis, utilizing structured templates and clear filtering principles.
[EXTERNAL_DOWNLOADS]: The skill integrates with multiple external data providers including crypto news outlets (Cointelegraph, Coindesk), macro financial sources (CNBC, Fed), and social media platforms (Weibo, GitHub) through specialized tools. These downloads are central to the skill's primary purpose.
[PROMPT_INJECTION]: Analyzed for indirect prompt injection risks given its function of processing external news content.
Ingestion points: External data is ingested through the news_feed, social_trending, and tradfi_news tools (SKILL.md).
Boundary markers: The skill uses structured output templates but does not implement specific 'ignore instructions' delimiters for the ingested content.
Capability inventory: The skill does not possess capabilities to execute arbitrary code, write to the filesystem, or perform network operations outside of the predefined data-fetching tools.
Sanitization: Content is processed through synthesis and summarization, which inherently reduces the risk of direct instruction execution from the source data.
[REMOTE_CODE_EXECUTION]: The skill references an external MCP server (https://datahub.noxiaohao.com/mcp) in a comment, which is used to provide the underlying tool capabilities for the agent environment.

news-briefing