managing-tilebox-jobs
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous examples for executing the
tileboxCLI tool to manage cloud resources, including submitting jobs (tilebox job submit), inspecting logs (tilebox job logs), and managing infrastructure (tilebox cluster). These are documented as the primary intended use of the skill for the vendor's own platform. - [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read job logs and span events using
tilebox job logsandtilebox job spans. While these sources could potentially contain untrusted data from task execution, this is an inherent surface area for any log-management tool. - Ingestion points: SKILL.md (via
tilebox job logs,tilebox job get,tilebox job list). - Boundary markers: None explicitly defined for log output.
- Capability inventory: Subprocess execution of the
tileboxCLI and a local Python virtual environment. - Sanitization: No specific sanitization or filtering of log content is mentioned.
Audit Metadata