python-bigquery-sdk
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No security issues or malicious patterns were identified. The skill correctly follows industry-standard best practices for interacting with Google Cloud services, including the use of Application Default Credentials and thread-safe client management.
- [INDIRECT_PROMPT_INJECTION]: The skill documents methods for querying and loading data from external sources (local files, Google Cloud Storage, and SQL results), which inherently defines an attack surface for indirect prompt injection. However, the skill provides specific and mandatory guidance on mitigating these risks.
- Ingestion points: The skill uses
client.query_and_wait,client.load_table_from_file, andclient.load_table_from_urias points where external data enters the environment. - Boundary markers: The documentation explicitly recommends using
QueryJobConfigandScalarQueryParameterto create clear boundaries between SQL commands and untrusted data. - Capability inventory: The skill utilizes network capabilities to communicate with Google Cloud APIs and file system access for data ingestion and export jobs.
- Sanitization: Parameterized queries are presented as mandatory for untrusted input, providing robust protection against SQL injection attacks.
Audit Metadata