The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from Lytics API responses (e.g., segment sizes and flow definitions) to manage customer journeys. This risk surface is mitigated by a mandatory confirmation gate requiring user approval for all mutation payloads, ensuring human review before any destructive action is taken (reference: references/confirmation-gate.md).
[DYNAMIC_EXECUTION]: The skill constructs curl commands dynamically for API interaction. Safety is ensured through specific instructions for URL-encoding path parameters using jq and using binary data flags for JSON payloads to prevent command injection (reference: references/api.md).

lytics-flows