The Agent Skills Directory

[SAFE]: The skill's logic is transparent and correctly performs its stated function without any malicious behavior.\n- [PROMPT_INJECTION]: The skill ingests external data (stock news headlines) which represents a surface for indirect prompt injection.\n
Ingestion points: The scripts/news.py script retrieves news content from the internet.\n
Boundary markers: The data is returned in a structured JSON format but lacks explicit delimiters for untrusted text fields.\n
Capability inventory: The skill's capabilities are limited to data retrieval and JSON output, with no high-risk command execution.\n
Sanitization: No sanitization of the external news content is performed before it is presented to the agent.

news-sentiment