The Agent Skills Directory

[SAFE]: The skill provides legitimate documentation and code templates for the aeon time series machine learning library.
[NO_CODE]: The skill consists entirely of documentation, references, and evaluation metadata. No executable scripts or binaries are distributed within the skill package.
[EXTERNAL_DOWNLOADS]: Fetches dataset archives from established research sources like the UCR/UEA Time Series Archives and the Monash Forecasting Archive via official library functions (e.g., load_classification, download_all_regression). These downloads target well-known research repositories.
[COMMAND_EXECUTION]: Uses uv pip install to manage the aeon package and uv run python to execute data analysis scripts. These are standard operations for the skill's intended purpose.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for external data files (.ts, .arff, .csv, .tsv).
Ingestion points: references/datasets_benchmarking.md lists various task-specific and format-specific loaders.
Boundary markers: None explicitly defined in the provided instructions.
Capability inventory: Shell execution (Bash) and file writing (write_to_ts_file) are available across the skill via the agent's tools.
Sanitization: Content is processed as numeric or symbolic data by specialized parsers in the aeon library, minimizing the risk of instruction injection.

alterlab-aeon