writing-sql
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructional examples for using Celigo CLI tools to discover database schemas and test query results. These include
celigo metadata types/fieldsfor schema discovery andceligo exports/imports invokefor triggering data operations. - [INDIRECT_PROMPT_INJECTION]: The skill defines the mechanism for interpolating untrusted data from records into SQL query templates.
- Ingestion points: Data from incoming records is injected into
rdbms.query,rdbms.once.query, andrdbms.query[]fields via Handlebars templates. - Boundary markers: The skill distinguishes between
{{ }}(for auto-quoted strings) and{{{ }}}(for raw identifiers and numbers) to manage data boundaries within SQL statements. - Capability inventory: The skill leverages database read and write capabilities using platform tools like
celigo exports invokeandceligo imports invoke. - Sanitization: The skill provides explicit safety instructions, explaining how the platform's Handlebars implementation automatically quotes and escapes values in double-braces to prevent SQL injection, and warns against manual quoting of triple-braced values.
Audit Metadata