The Agent Skills Directory

[SAFE]: The skill focuses on documenting the legitimate anndata Python library. All code samples represent standard, benign usage of the library for data science and bioinformatics tasks.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install standard Python packages (anndata, scanpy) using the uv tool. It also provides examples for accessing remote data via fsspec and urllib, which include explicit security best practices such as host and protocol validation to prevent unauthorized downloads.
[COMMAND_EXECUTION]: The skill uses the bash tool solely for environment setup and package management (e.g., uv pip install). These operations are transparent and consistent with the skill's purpose as a developer tool.
[DATA_EXFILTRATION]: No patterns of unauthorized data collection or exfiltration were found. File read and write operations are confined to expected data formats like .h5ad, .zarr, and .csv used in scientific workflows.
[PROMPT_INJECTION]: The instructions do not contain any attempts to override system prompts, bypass safety filters, or extract sensitive model information.

anndata