skills/jeremylongshore/tons-of-skills-marketplace/clickhouse-migration-deep-dive/Gen Agent Trust Hub
clickhouse-migration-deep-dive
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill facilitates the creation and execution of a migration runner (
runner.ts) and SQL files. This workflow involves generating executable code and running it via thenodeenvironment, which is a common but sensitive pattern in automated maintenance tasks. - [INDIRECT_PROMPT_INJECTION]: The migration runner executes SQL content directly from generated files. If the agent generates these migrations based on external or untrusted schema specifications, it could be susceptible to indirect SQL injection where malicious instructions are embedded in the migration logic.
- Ingestion points: SQL migration files located in
migrations/sql/generated by the agent. - Boundary markers: None identified; the runner executes the statements sequentially.
- Capability inventory:
client.commandfor database operations andBashfor executing the TypeScript runner. - Sanitization: The runner lacks explicit validation or sanitization of the SQL statements before execution.
Audit Metadata