The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The documentation and code examples in the skill recommend using the Python pickle module for caching embeddings and model outputs.
Evidence: embeddings = pickle.load(f) found in SKILL.md and references/examples.md.
Risk: The pickle module is not secure. Loading untrusted or modified pickle data can lead to arbitrary code execution on the user's system.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of the molfeat library and its various extras from PyPI, as well as pre-trained chemical models from external repositories.
Evidence: uv pip install molfeat and its variants for DGL, Graphormer, and Transformer models.
Evidence: Usage of PretrainedMolTransformer which fetches model weights from HuggingFace and other model hubs.
[COMMAND_EXECUTION]: The skill provides explicit shell commands for the agent or user to execute for environment setup.
Evidence: Installation instructions using uv pip install in SKILL.md and references/examples.md.
[NO_CODE]: This skill is currently composed entirely of Markdown instruction and reference files. It does not contain any standalone scripts (.py, .js, .sh) or binary executables.

molfeat