The Agent Skills Directory

[SAFE]: The skill exclusively references official Grafana Cloud domains (grafana.com, grafana.net) for documentation and API interactions, which are trusted vendor resources.
[SAFE]: Sensitive information such as API keys and endpoints are handled via environment variable lookups (sys.env) or designated placeholders (e.g., <token>, <stack>), following security best practices for credential management.
[COMMAND_EXECUTION]: Contains standard administrative commands using curl and jq to fetch optimization recommendations from official Grafana APIs and to perform local reloads of the Alloy service. These operations are transparent and restricted to the local environment or vendor-owned infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill includes PromQL and LogQL queries that process telemetry data. While this introduces an ingestion surface for potentially untrusted data (logs/metrics), the skill does not chain this data with dangerous capabilities or autonomous decision-making that would enable exploitation.

cost-management