provision-failure-analysis
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references downloading CI artifacts and logs from Google Cloud Storage (https://storage.googleapis.com/). This is a well-known service used for hosting publicly accessible build logs from OpenShift Prow jobs, which are vendor-controlled resources in this context.\n- [COMMAND_EXECUTION]: The skill instructs the agent to analyze pipeline configurations, Shell scripts, and Helm releases within the codebase. This involves inspecting logic in files like
tooling/templatize/pkg/pipeline/run.goto understand execution flow and failure paths.\n- [PROMPT_INJECTION]: The skill processes external data from build logs and cluster logs, which constitutes an indirect prompt injection surface.\n - Ingestion points: Build logs retrieved from Google Storage and various Kusto tables (
containerLogs,kubernetesEvents,kubeAudit,systemdLogs,aksEvents) in theServiceLogsdatabase.\n - Boundary markers: The skill does not define specific delimiters or instructions for the agent to distinguish between its instructions and potentially malicious content embedded within the log data.\n
- Capability inventory: The skill allows the agent to perform batch analysis, write data to a local directory (
zz-ignore/provision-failures/), and recommend infrastructure changes based on the data findings.\n - Sanitization: There are no explicit instructions for filtering or sanitizing log content to prevent instructions in logs from influencing agent behavior.
Audit Metadata