The Agent Skills Directory

[SAFE]: The skill uses established bioinformatics packages such as omicverse, scanpy, and scvelo for developmental trajectory interpolation. These tools are used for their intended purpose in a scientific workflow.
[EXTERNAL_DOWNLOADS]: The skill references the scv.datasets.dentategyrus() function, which loads a public reference dataset from the scvelo library's official repository. This is a standard procedure for bioinformatics benchmarking and development.
[COMMAND_EXECUTION]: The skill instructions involve executing model training (beta-VAE and GNN) and data processing steps. These operations are performed using the vendor's omicverse library and the pyVIA package, which are consistent with the skill's stated goal of trajectory interpolation.
[INDIRECT_PROMPT_INJECTION]: The skill processes external biological data (bulk RNA-seq and gene mappings). * Ingestion points: Data is loaded from local files and public datasets via ov.utils.read and scv.datasets.dentategyrus (SKILL.md, reference.md). * Boundary markers: There are no explicit markers to distinguish data content from instructions. * Capability inventory: The skill has the capability to write to the local filesystem using adata.write_h5ad and saving model checkpoints. * Sanitization: The skill includes gene ID harmonization and defensive validation steps (e.g., asserting column existence and gene overlap) to ensure data integrity before processing.

bulktrajblend-trajectory-interpolation