game-testing-first-day-report
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands using the
ae-cliutility. These commands are used for project metadata retrieval, metadata export to temporary files, and running analytical SQL queries via parallel bash calls. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data exported from external systems into temporary files (
/tmp/events.json,/tmp/props.json). This constitutes a potential attack surface where malicious data in the analytics platform could attempt to influence the agent's report generation process. The evidence chain includes ingestion points in the/tmp/directory and a capability inventory involving shell and python execution. The skill provides basic sanitization guidance, such as double-quoting Trino identifiers. - [DYNAMIC_EXECUTION]: The skill dynamically generates SQL queries and JSON definitions at runtime based on user-provided parameters like dates and project IDs. These are then executed via the CLI tool. This is a core functional requirement for the data analysis tasks it performs.
Audit Metadata