The Agent Skills Directory

[SAFE]: The skill consists of documentation and code examples for using the UMAP library for dimensionality reduction. No malicious patterns were identified across the analyzed files.- [EXTERNAL_DOWNLOADS]: The documentation provides standard installation instructions using uv pip install. It references well-known and trusted libraries including umap-learn, scikit-learn, matplotlib, hdbscan, and tensorflow.- [COMMAND_EXECUTION]: Shell commands are provided for installing the necessary Python packages. These are routine tasks for setting up a development environment and do not involve suspicious or unauthorized actions.- [REMOTE_CODE_EXECUTION]: No patterns of remote code execution, such as fetching and executing arbitrary scripts from untrusted sources, were found.- [DATA_EXFILTRATION]: No evidence of unauthorized data access or exfiltration to external domains was detected.- [PROMPT_INJECTION]: The instructions are focused on providing technical guidance for using the UMAP library and do not contain attempts to subvert agent behavior or bypass safety guardrails.- [INDIRECT_PROMPT_INJECTION]: The skill provides examples for processing data. The ingestion points are local data variables, and the capabilities are limited to standard data transformation methods within the UMAP and scikit-learn libraries. Sanitization is not applicable for these static documentation examples.

umap-learn