define-slo-sli-sla
Fail
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads an external binary from a third-party GitHub repository (
slok/sloth) using thewgetcommand.\n- [REMOTE_CODE_EXECUTION]: The procedure provides instructions to grant execution permissions to the downloaded binary (chmod +x) and execute it immediately to generate configuration files.\n- [COMMAND_EXECUTION]: The skill usessudoto move the downloaded binary into/usr/local/bin, which constitutes privilege escalation to modify system-level directories.\n- [COMMAND_EXECUTION]: Automated shell commands are used to trigger service reloads for a local Prometheus instance usingcurlagainstlocalhost:9090.\n- [PROMPT_INJECTION]: Vulnerability to indirect prompt injection identified through the following surface:\n - Ingestion points: Data enters the agent context from local YAML configuration files and the Prometheus API (
http://prometheus:9090/api/v1/query).\n - Boundary markers: Absent; there are no delimiters or instructions to ignore embedded commands within the ingested metrics or configuration data.\n
- Capability inventory: The skill has access to powerful administrative tools including
Bash,Write,Edit,Grep, andGlob.\n - Sanitization: No evidence of input validation or escaping for the external content retrieved from the metrics provider or local files.
Recommendations
- AI detected serious security threats
Audit Metadata