The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the arboreto library and its common data science dependencies (including pandas, numpy, dask, and scikit-learn) using the uv package manager. These are standard installations from public registries.
[REMOTE_CODE_EXECUTION]: The skill includes documentation and code for connecting to remote Dask schedulers via TCP (e.g., tcp://scheduler:8786) to facilitate distributed computation. While this mechanism involves remote task execution, it is the primary intended functionality of the Dask integration for processing large-scale datasets.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of processing external data files.
Ingestion points: In scripts/basic_grn_inference.py, the skill reads gene expression data from a user-specified TSV file using pd.read_csv.
Boundary markers: The instructions do not define boundary markers or provide the agent with guidance to ignore potential natural language instructions embedded within the gene expression matrices.
Capability inventory: The skill possesses the capability to execute Python code and manage multi-process tasks via Dask and the distributed library as seen in SKILL.md and scripts/basic_grn_inference.py.
Sanitization: The script performs standard numeric data loading via Pandas but lacks specific sanitization or filtering to prevent the processing of malicious payloads embedded in the data.

arboreto