ax-prompt-builder

Assemble the AI agent's system prompt by ordering composable modules to construct the final prompt structure. This approach allows you to conditionally include or drop modules based on the current context and available token budget.

Handle token budgeting for a prompt system by conditionally including or dropping modules based on the available context. This ensures the assembled system prompt stays within the required token limits while preserving core agent functionality.

Context-aware tool filtering works by generating a ToolFilterContext during prompt assembly. This context evaluates the current task requirements to dynamically select which tools the AI agent can access.

Manage bootstrap mode in a prompt system to handle first-run identity discovery and graceful degradation. This allows the AI agent to initialize its core behaviors safely when encountering an unknown environment.

Configure runner selection guidance to enable task delegation to sub-agents within the prompt system. This provides the AI agent with the necessary instructions to route and execute delegated tasks effectively.

An AI agent drops prompt modules during execution due to conditional inclusion rules triggered by token budgeting constraints. Modules are dynamically omitted to maintain operational stability and fit the context window.