The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute gcx CLI commands. These commands are restricted to read-only operations such as list, get, query, and status to showcase Grafana Cloud resources.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, including log entries, alert annotations, and AI-generated narratives from Grafana Assistant investigations. While this data could theoretically contain instructions intended to influence the agent, the risk is mitigated by the skill's read-only scope and the nature of the demo context.
Ingestion points: Data enters the context via gcx logs query, gcx alert rules list, and gcx assistant investigations get-narrative outputs.
Boundary markers: None explicitly defined in the provided instructions for separating command output from narration.
Capability inventory: The agent has access to Bash for command execution and AskUserQuestion for user interaction.
Sanitization: No specific sanitization or filtering of the CLI output is performed before presentation to the user.
[SAFE]: The skill identifies as being authored by Grafana and interacts exclusively with Grafana's official gcx CLI tool. All operations are aligned with the intended purpose of providing a read-only product demonstration.

gcx-demo