The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes alert titles, summaries, and labels fetched from Grafana OnCall, which presents a surface for indirect prompt injection. Maliciously crafted data in an alert could potentially attempt to override agent instructions or influence its decision-making process.
Ingestion points: The agent retrieves alert data via the gcx irm oncall alert-groups list and gcx irm oncall alert-groups get commands as defined in SKILL.md.
Boundary markers: There are no specific instructions or delimiters provided to the agent to segregate external alert content from its primary instructions.
Capability inventory: The skill is granted access to the Bash tool and can execute state-altering commands such as acknowledge, silence, resolve, and delete through the gcx CLI.
Sanitization: The skill does not implement filtering or sanitization of the alert data before it is processed by the agent context.
[COMMAND_EXECUTION]: The skill instructions utilize the Bash tool to run the gcx CLI and jq for data processing and alert management. This is the intended functionality of the skill for interacting with the Grafana OnCall service.

oncall-triage