The Agent Skills Directory

[SAFE]: The skill provides instructions for querying telemetry data using the cx CLI, a legitimate vendor-owned utility. All described operations are focused on data retrieval and analysis, with an explicit mention that query commands are read-only.\n- [PROMPT_INJECTION]: The skill involves processing external telemetry data (logs, spans, and RUM events), which represents a surface for indirect prompt injection. This is documented for awareness but is consistent with the skill's purpose. Evidence: 1. Ingestion points: Output from cx logs, cx spans, and cx metrics. 2. Boundary markers: No specific delimiters or 'ignore' instructions are provided in the current templates. 3. Capability inventory: Shell command execution via the cx tool. 4. Sanitization: No explicit sanitization of telemetry content is performed.

cx-telemetry-querying