Right now the agent is not able to get any meaningful information from images or GIFs.

• Add png and gif image description using moondream (see other issue)
• Integrate with search client, interaction client and feed client so the agent has the image described as context