The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation instructs users to install several reputable Python packages from public registries, including the Vaex meta-package and its components, as well as cloud-native filesystem backends such as s3fs, gcsfs, and adlfs.\n- [DYNAMIC_EXECUTION]: The documentation describes the use of Numba-based Just-In-Time (JIT) compilation and the 'apply' method for executing user-defined Python logic. These are documented as standard performance optimization features for large-scale data processing.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection because it is designed to ingest and process data from external, untrusted sources such as HDF5/CSV/Arrow/Parquet files, cloud storage (S3, GCS, Azure), and SQL databases.\n
Ingestion points: Data entry occurs via vaex.open(), vaex.from_csv(), and various remote storage connectors mentioned in references/io_operations.md.\n
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat data content as untrusted or to ignore embedded natural language commands.\n
Capability inventory: The skill enables interaction with the local system through tools like Bash, Write, and Edit, and provides capabilities for network-based data access.\n
Sanitization: The skill does not include specific mechanisms for sanitizing or validating the content of the tabular data it processes.

vaex