The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the gcx command-line tool to interact with Grafana APIs. The instructions detail specific subcommands for fetching state (gcx dashboards get), managing resources (gcx resources push/pull/delete), and generating snapshots (gcx dashboards snapshot). These operations are conducted using the agent's shell environment to interact with the user's configured Grafana instance.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it retrieves and processes dashboard definitions in JSON and YAML formats from remote Grafana instances.
Ingestion points: Dashboard content is ingested via gcx dashboards get and gcx resources pull commands as described in SKILL.md and references/resource-operations.md.
Boundary markers: The instructions do not specify explicit delimiters or "ignore" instructions for the agent when processing dashboard content.
Capability inventory: The skill has significant write and delete capabilities, including gcx dashboards update, gcx resources push, and gcx resources delete.
Sanitization: There is no explicit sanitization step mentioned for the query strings or metadata contained within the retrieved dashboard resources.
Assessment: While this ingestion surface exists, it is the primary intended purpose of a dashboard management tool. Since the tool is authored by the vendor (Grafana) for administrative use, this is considered a standard functional characteristic rather than a malicious pattern.
[SAFE]: All external references and tools trace back to the official Grafana vendor ecosystem (gcx CLI, grafana.app annotations). The skill includes safety preflights like gcx config check and enforces manager protection boundaries to prevent accidental overwrites of resources managed by other tools.

manage-dashboards