agent-orchestrator

Minting unique operational IDs for pipeline work requests involves generating a unique UUID for each incoming request. This ensures every work item is distinctly tracked and managed before it advances to the Planner stage.

Rule-based classification for intake pipelines determines the operational mode and execution type by evaluating predefined rules against the incoming request. It validates repository data and automatically categorizes the intent for proper routing.

The best way to structure user intent before sending it to a Planner is to package it into a normalized intake packet. This packet includes a minted operational ID, classified execution type, and validated repository context for immediate processing.

Yes, you can validate incoming work requests against existing repository registries. The intake process checks the submitted user intent and inbox JSON data against these registries to ensure requests target valid repositories.

You do not need to manually specify an execution type to classify work requests. The intake normalization process automatically determines the execution type based on predefined rules applied to the parsed user intent.