The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables the execution of ClickHouse SQL queries through the chdb.query() and Session.query() APIs. This is the primary function of the skill, allowing for complex data analysis, joins across disparate sources, and file system interactions (reading/writing via file() and Session persistence).
[EXTERNAL_DOWNLOADS]: The skill documents the use of ClickHouse table functions such as url(), s3(), gcs(), and azureBlobStorage() to fetch and query data from remote URLs and cloud storage providers. These are standard features of the documented library.
[DYNAMIC_EXECUTION]: The references/api-reference.md file describes the User-Defined Function (UDF) capability via the @chdb_udf() decorator. This allows users to register Python functions that are executed dynamically within the SQL engine's process during query execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, potentially untrusted sources (e.g., remote CSV files or databases).
Ingestion points: Data enters the environment via file(), url(), s3(), and various database connector table functions listed in references/table-functions.md.
Boundary markers: The documentation does not specify the use of prompt-level boundary markers for data processed via SQL.
Capability inventory: The library can execute complex SQL, read/write local files (via Session paths or file() outputs), and perform network operations to fetch remote data.
Sanitization: No specific sanitization or filtering of external data content is described, though standard ClickHouse SQL parsing provides a layer of structure.

chdb-sql