The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill scripts download the @toolbox-sdk/server@1.1.0 package from the NPM registry at runtime using the npx command to provide core analytical functionality.
[COMMAND_EXECUTION]: The skill invokes its analytical tools as subprocesses using child_process.spawn. It includes robust logic for handling shell execution on Windows and Unix systems, including appropriate argument escaping.
[PROMPT_INJECTION]: The skill exposes surfaces for indirect prompt injection as part of its intended functionality for processing complex data queries.
Ingestion points: Parameters such as input_data, user_query_with_context, and history_data in SKILL.md accept SQL queries and natural language prompts that represent untrusted data sources.
Boundary markers: The instructions do not define explicit boundary markers to separate untrusted data from system instructions.
Capability inventory: The skill possesses capabilities for data retrieval and analysis through the integrated toolbox and BigQuery connectors.
Sanitization: While the scripts perform shell-level escaping of arguments, the skill does not implement specific sanitization or filtering for the content of processed SQL queries or prompts.

bigquery-analytics