The scaffolding around a model that runs the agent loop: orchestration, tool dispatch, error handling, memory between steps, stop conditions. Most agent failures are harness failures.
Definition: The scaffolding around a model that runs the agent loop: orchestration, tool dispatch, error handling, memory between steps, stop conditions. Most agent failures are harness failures.
The scaffolding around the model that runs the agent loop: orchestration, tool dispatch, error handling, memory between steps, stop conditions. The model is the engine; the harness is the chassis. Most agent failures are harness failures, not model failures.
A system that uses a language model to drive multi-step work in a loop: reason, call a tool, observe the result, decide the next step. Distinct from single prompt-and-response interactions.
A model capability for invoking external functions: web search, database queries, code execution, API calls. Required ingredient for any non-trivial agent.
Open standard for connecting AI models to external tools and data sources through a shared interface. Supported by multiple AI vendors, replacing one-off integrations with a portable contract.