> ## Documentation Index
> Fetch the complete documentation index at: https://vokra.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Agents

> The reusable model, instructions, and tool permissions behind a job.

An agent is a reusable worker inside one workspace. It defines a name and description, system instructions, an LLM credential and model, runtime settings, and the connected MCP servers it may use.

Agents do not own schedules or webhooks. Create a [job](/concepts/jobs) for an agent, then add the trigger to that job. Multiple jobs can reuse the same agent with different prompts.

Disabling an agent prevents its jobs from running. Keep an agent narrow enough that its instructions and tool permissions are easy to review.