byted-bytehouse-data-quality-inspector
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The shell script
scripts/export_config.shuses theevalcommand to load environment variables parsed from a local JSON configuration file. While the script usesjqwith the@shfilter to properly escape values, the use ofevalon generated strings remains a sensitive execution pattern. - [PROMPT_INJECTION]: The skill processes external data (database schema and table data) to generate quality reports, creating a surface for indirect prompt injection where malicious database content could influence agent behavior.
- Ingestion points: Schema metadata from
system.columnsand table content viascripts/inspector.py. - Boundary markers: None; the script does not use specific delimiters or instructions to ignore embedded commands in the data.
- Capability inventory: Read-only database access via
clickhouse-connectfor executing statistical queries. - Sanitization: Database and table identifiers are interpolated into SQL queries using f-strings without additional sanitization or escaping.
Audit Metadata