The Agent Skills Directory

[SAFE]: The skill documentation describes standard and legitimate use cases for the CARTO CLI tool, consistent with the official functionality of the vendor (CartoDB).- [COMMAND_EXECUTION]: The skill utilizes the carto CLI for warehouse operations such as import, activity export, and sql job. These are the primary intended functions of the skill and do not involve unauthorized command execution.- [EXTERNAL_DOWNLOADS]: The carto import --url command facilitates fetching data from remote URLs. This is a core feature for moving geospatial data into a warehouse and does not involve downloading or executing arbitrary scripts in the local environment.- [PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection through its data ingestion capabilities.
Ingestion points: External data is ingested via local files (--file) and remote URLs (--url) as documented in SKILL.md and references/imports.md.
Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore embedded instructions in the imported data.
Capability inventory: The skill allows the execution of the carto CLI tool and the submission of SQL jobs to a connected data warehouse, as described in references/tilesets.md.
Sanitization: Data validation and sanitization are handled by the CARTO CLI and backend services during the parsing and loading process.

carto-import-export-data