The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external New Relic telemetry sources which may contain attacker-controlled content such as error messages or log entries. This creates a surface for indirect prompt injection if the agent interprets these strings as instructions.
Ingestion points: Results from newrelic nrql query in nrql/SKILL.md and apm/SKILL.md.
Boundary markers: The instructions do not provide explicit delimiters or warnings for the agent to ignore instructions embedded in the telemetry data.
Capability inventory: The skill uses the newrelic CLI and jq for data retrieval and processing.
Sanitization: While the skill includes NRQL input escaping in its helper scripts (referenced in local.patch), it does not sanitize the output of queries for prompt injection patterns.
[COMMAND_EXECUTION]: The skill provides numerous examples of newrelic CLI usage for querying and managing observability data. These are standard operational commands for the tool's intended purpose.
[EXTERNAL_DOWNLOADS]: The skill references downloading the newrelic CLI from its official GitHub releases page, which is a well-known and trusted source for this utility.

newrelic-cli-skills