The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's primary function is to execute the gcx command-line utility. It performs a wide range of operations including querying telemetry data, managing dashboards, and modifying alert rules. It explicitly instructs the agent to discover commands via help flags and tree structures to ensure accuracy.
[DYNAMIC_EXECUTION]: The skill implements a dynamic context loading pattern where it instructs the agent to retrieve task-specific instructions directly from the gcx binary using gcx agent skills get <name> -otext. This ensures that the instructions the agent follows are always compatible with the version of the tool installed on the system.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources, creating a surface for indirect prompt injection.
- Ingestion points: Commands read telemetry data (logs, metrics, traces), dashboard manifests, and large output results stored in temporary "spill files" (gcx-results-*.json). It also ingests instructions from the gcx agent skills command.
- Boundary markers: The skill does not define specific prompt delimiters for external content within this file, though it suggests using structured output flags (--json, --jq).
- Capability inventory: The agent has the capability to execute shell commands (gcx), read/write files via the CLI, and perform network operations through the managed Grafana API.
- Sanitization: The skill relies on built-in CLI flags like --jq for data transformation and gcx resources validate to check the integrity of manifests before they are pushed.
[EXTERNAL_DOWNLOADS]: The skill documentation guides the user to install the core dependency using brew install gcx, which is a standard package management operation.

use-gcx