The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from an external monitoring platform, creating a potential surface for indirect prompt injection if the retrieved content is untrusted.
Ingestion points: The tools query_metrics, get_metric_metadata, and get_dashboard in SKILL.md fetch metric names, metadata, and dashboard configurations from the Datadog API.
Boundary markers: The instructions do not define specific delimiters or warnings for the agent to ignore embedded instructions in the retrieved data.
Capability inventory: The skill is scoped to metric and dashboard operations via the datadog-mcp transport. It does not possess capabilities for local file system writes, arbitrary shell command execution, or other high-privilege actions.
Sanitization: No explicit validation or filtering logic is specified for the external data before it is presented to the agent.
[EXTERNAL_DOWNLOADS]: The skill facilitates communication with Datadog's official monitoring services.
Evidence: The skill connects to the datadog-mcp server and defaults to the well-known datadoghq.com site for fetching metrics and dashboard definitions.

datadog-metrics