The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests existing SLO definitions and configuration from the Grafana environment to perform updates and GitOps synchronization. If a malicious actor embeds instructions within SLO metadata fields (such as names, descriptions, or alert annotations) in the Grafana instance, these instructions could influence the agent during the YAML processing phase.
Ingestion points: SKILL.md (Workflows 2 and 3) utilizes gcx slo definitions get and gcx slo definitions pull to import data from Grafana APIs into the agent's context.
Boundary markers: The skill does not employ specific boundary markers or "ignore instructions" headers when processing the imported YAML content.
Capability inventory: The skill has access to the Bash tool for executing CLI commands (including network-capable gcx push/delete operations) and the Write tool for file modifications.
Sanitization: The skill relies on structural YAML formatting but does not explicitly sanitize string fields for embedded agent instructions before processing.

slo-manage