databricks-performance-tuning
Pass
Audited by Gen Agent Trust Hub on Apr 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Multiple functions in
SKILL.md, includingoptimize_delta_table,enable_liquid_clustering, andenable_predictive_optimization, use f-string interpolation to build SQL commands. This creates a risk of SQL injection where a malicious table name could be used to execute arbitrary SQL statements on the Spark cluster.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes external data (table and column names) and interpolates it into executable code without sanitization.\n - Ingestion points: Function parameters
table_name,z_order_columns, andcluster_columnswithinSKILL.md.\n - Boundary markers: Absent; there are no delimiters or warnings to the agent to ignore instructions embedded within the data.\n
- Capability inventory: Uses
spark.sql()to execute commands on the data platform (identified inSKILL.md).\n - Sanitization: Absent; input is directly concatenated into SQL strings without escaping or validation.
Audit Metadata