Implementing Guardrails in an Automated SDR Flow - Medium

Quoted on this wiki

Every place a page here uses this source, in the order the words come in it.

1. Guardrail → A protective check on inputs/outputs before risky actions (e.g., detect names). “Trace → Adds observability to a run.” Code walkthrough

   In AI SDR agent guardrails

2. Agents → LLM-powered workers with instructions (e.g., “write a witty email”). “Tools → Functions or agents exposed as callable actions (e.g., send an email).” Runner → Executes an agent with a given message, coordinating tool calls.

   In Tools and AI