The Agent Skills Directory

[SAFE] (SAFE): No security risks or malicious patterns were identified in the analysis of the instructional content. The skill defines a persona and best practices for risk management without including scripts or configuration that could interact with the system host or external networks.\n- [NO_CODE] (INFO): This is a prompt-based skill that does not contain Python scripts, Node.js packages, or shell commands. The risk of runtime exploitation (RCE, privilege escalation) is negligible as no execution environment is targeted.\n- [INDIRECT_PROMPT_INJECTION] (INFO): The skill processes user-provided financial data (trades, portfolio positions). While it lacks explicit boundary markers for this data, it also lacks any tools or capabilities (network, file write) that could be used to exfiltrate data or cause side effects if an injection were to occur in the input data stream.

risk-manager