datahub-quality
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied strings for assertion descriptions, incident titles, and SQL code, which are interpolated into CLI commands. Ingestion points: Assertion parameters and incident metadata in SKILL.md. Boundary markers: Explicit 'Anti-injection rule' instructing the agent to ignore user-supplied instructions. Capability inventory: Uses Bash(datahub *) and temporary file writes for GraphQL operations. Sanitization: Direct instructions to reject shell metacharacters in CLI arguments.
- [COMMAND_EXECUTION]: The skill invokes the datahub CLI tool. It includes safety checks requiring the agent to validate URN formats and strip shell metacharacters from user-provided inputs to prevent command injection.
- [DYNAMIC_EXECUTION]: The skill generates temporary GraphQL query files in /tmp to manage long query strings that would otherwise exceed shell command line limits, representing a dynamic file creation and execution pattern.
Audit Metadata