Plugin agentic-ai app to run images thru multi modal model

agentic ai examples

scenario: hey rescuebox can you take these images and find me details about teenagers with details
steps would involve a workflow :

1. first pick from existing plugins a scan of images to get a caption or find images with faces
2. second from this filtered list of images , run thru another plugin to detect age like teenagers
3. third from this subset run thru image summarize plugin to get details about the images

these 3 plugins are selected and sequenced by an agentic-ai app-plugin the goal of this effort in rescuebox

this can become complex and have dependencies on other features like pipeline capability
pipeline

Learn from chatgpt about these four levels of Agent maturity including (1) large language model (LLM)-powered information retrieval agents, (2) single-task agentic workflows, (3) cross-system agentic workflow orchestration, and (4) multi-agent constellations. Specific to this issue level 2 and 3 is the focus.

Start by reviewing the linked agentic AI examples and the pipeline issue (#56), then identify how existing plugins are selected and sequenced. The intended result is a workflow that filters images, detects teenagers, and summarizes the resulting subset, but the issue does not name implementation files or tests.