The Agent Skills Directory

[SAFE]: The skill is informational in nature, consisting only of markdown text and reference documents. No executable scripts, tool configurations, or privilege escalations are present.
[EXTERNAL_DOWNLOADS]: The skill references a wide array of external URLs in references/article-index.md to ground its financial analysis. These sources include reputable organizations such as CNBC, NPR, and Yale University. These are intended for data retrieval for educational purposes and do not involve remote code execution.
[PROMPT_INJECTION]: The skill provides a framework for analyzing user-supplied market scenarios, which constitutes an ingestion point for untrusted data. However, the skill lacks any capabilities to perform dangerous operations like writing to the file system, executing subprocesses, or making network requests.
Ingestion points: User-provided market questions and scenarios processed by the Druckenmiller framework in SKILL.md.
Boundary markers: None present; the instructions do not include delimiters to isolate user input.
Capability inventory: No tools are defined in the frontmatter, and no scripts are included in the skill.
Sanitization: None observed; the skill does not filter or sanitize user input before interpolation into the analysis framework.

investor-stanley-druckenmiller