Capability

A Capability is a self-contained unit of agent behavior that exposes its features as tools and prompt sections. Any LLM backend that supports tool calling can drive it.

The Capability interface follows the Component pattern: each feature (flows, triage, extraction, handoffs, guardrails) is a capability that plugs into a CapabilityHost. The host collects tools, builds prompts, and routes tool results — the LLM backend is irrelevant.

Methods

getPromptSections()

getPromptSections(): CapabilityPromptSection[];

What does this capability contribute to the system prompt?

Returns

CapabilityPromptSection[]

---

getTools()

getTools(): ToolDeclaration<unknown, unknown>[];

What tools does this capability expose for the current state?

Returns

ToolDeclaration<unknown, unknown>[]

---

processToolResult()

processToolResult(
   toolName,
   args,
   result): CapabilityAction | null;

A tool was called and executed. Does this capability handle the result?

Parameters

Parameter	Type	Description
toolName	string	The name of the tool that was called
args	unknown	The arguments passed to the tool
result	unknown	The result returned by the tool’s execute()

Returns

CapabilityAction | null

An action telling the host what to do, or null if this capability doesn’t handle this tool.