The Agent Skills Directory

[METADATA_POISONING]: The skill's metadata in SKILL.md identifies the author as 'K-Dense Inc.', which contradicts the platform-provided author identity 'leonchaox'. This inconsistency is misleading regarding the provenance of the skill.
[DYNAMIC_EXECUTION]: The skill documents and promotes the use of numba.jit (JIT compilation) within references/performance.md to compile custom functions into machine code at runtime for optimization. This pattern of dynamic execution introduces risks if user-influenced logic is passed to the compiler.
[INDIRECT_PROMPT_INJECTION]: The skill is primarily designed to ingest and manipulate large tabular datasets from diverse and untrusted external sources, creating a potential attack surface.
Ingestion points: Data is loaded via vaex.open(), vaex.from_csv(), and cloud storage protocols (s3://, gs://, az://) as documented in references/core_dataframes.md and references/io_operations.md.
Boundary markers: The instructions do not specify any delimiters or warnings to the agent regarding potential instructions embedded within the processed data.
Capability inventory: The skill includes functions for file system writes (export_hdf5, export_csv) and network communication for remote data access and server connections (ws://), as detailed in references/io_operations.md.
Sanitization: There are no mentioned mechanisms for validating, filtering, or sanitizing the content of the external datasets before they are processed by the agent.

vaex