The Agent Skills Directory

[SAFE]: No malicious patterns or suspicious activities were detected. The skill is designed as a structured reasoning framework rather than an executable script.
[SAFE]: Explicit guardrails are provided to prevent the AI from impersonating public figures or treating their opinions as authoritative proof, emphasizing a 'framework-not-persona' approach.
[SAFE]: The skill incorporates comprehensive safety checks that reject any output proposing unethical, illegal, or privacy-invasive shortcuts, and explicitly mandates human safety and quality guardrails.
[SAFE]: File system operations are limited to a specific, scoped directory (.hypercore/elon-musk/) for the purpose of saving and resuming analysis reports.
[PROMPT_INJECTION]: The skill processes user-provided problem descriptions, which represents a surface for indirect prompt injection. However, the instructions mandate a rigorous classification and deconstruction workflow (A/B/C matrix) that treats the input as data to be analyzed, effectively mitigating the risk of the input being executed as instructions.

elon-musk