The Agent Skills Directory

[SAFE]: The skill is authored by 'grafana' and uses the vendor's specialized CLI tool (gcx) for monitoring and analysis. All operations are consistent with the stated purpose of SLO optimization.
[COMMAND_EXECUTION]: The skill uses Bash to run gcx commands for retrieving SLO definitions and metrics. Parameters like UUIDs and datasource UIDs are retrieved from the system or user. The skill's logic is purely advisory, which significantly minimizes the risk and impact of potential command injection.
[DATA_EXFILTRATION]: There is no evidence of sensitive data access or transmission to unauthorized domains. Activity is restricted to the user's configured Grafana instance via the gcx tool.
[PROMPT_INJECTION]: The skill processes metric data and SLO definitions which may contain untrusted strings. While this represents a theoretical indirect prompt injection surface, the skill focuses on numerical values and statistical analysis, and provides only advisory text to the user, reducing the likelihood of malicious instruction execution.

slo-optimize