ProtectedactionVerb used in error messages.
ProtectedentityEntity searched by this action.
ProtectedincludeWhether each result includes sub-agent details + default artifact type (candidate variant).
ProtectedapplyVariant-specific invocation/parent filtering. Default excludes agents whose invocation mode is 'Sub-Agent' (they're meant to be called by other agents).
ProtectedbuildBuilds the per-agent result object, including sub-agent details for the candidate variant.
ProtectedgetAction names available to the given agent.
ProtectedgetSub-agents (direct children + relationship-linked) for the given agent.
ProtectedInternalInternal method that must be implemented by derived action classes. This is where the actual action logic should be implemented.
The action execution parameters including typed context
Promise resolving to the action result
Executes the action with the provided parameters.
The action execution parameters including context
Promise resolving to the action result
Finds candidate AI agents for a given task using semantic search.
Backed by the unified
Provider.SearchEntitypipeline (semantic mode over the daily-synced "AI Agents Search" EntityDocument). Extends the shared agent base (see BaseFindAgentsAction) to additionally surface each agent's sub-agents and default artifact type, and to optionally exclude Sub-Agents and child agents via theExcludeSubAgentsparameter.Inputs:
TaskDescription(required),MaxResults(default 5),MinimumSimilarityScore(cosine 0-1, default 0.5),IncludeInactive(default false),ExcludeSubAgents(default true).Example