Self-improving AI Employees for everyone
Python ai-agentsai-employeeai-employeesassistantautonomous-agentsco-employeescompanygrok-botinstinctloop-agentsmuseown-your-datapersonalself-hosted
docs(nodeSpec): clarify reactivity expectations for getCachedNodeSpec
trohitga8b2f3f1999b
diff --git a/client/src/lib/nodeSpec.ts b/client/src/lib/nodeSpec.tsindex 0cdac5b9..6974a722 100644--- a/client/src/lib/nodeSpec.ts+++ b/client/src/lib/nodeSpec.ts@@ -63,6 +63,13 @@ export async function prefetchAllNodeSpecs(sendRequest: SendRequest): Promise<vo* Synchronous read of a previously-cached NodeSpec. Used by flag-gated* render paths that need a description on-demand without suspending.* Returns `null` if not yet fetched.+ *+ * Reactivity: Dashboard's `prefetchAllNodeSpecs` warms every spec on+ * WebSocket connect, before any node mounts. Consumer components call+ * this on every render; React Flow re-renders nodes on viewport /+ * selection / parameter changes, which gives the spec a chance to+ * arrive. If you need eager reactivity, wrap with `useQuery` using+ * the shared `nodeSpecQueryKey(type)`.*/export function getCachedNodeSpec(nodeType: string): NodeSpec | null {return queryClient.getQueryData<NodeSpec | null>(nodeSpecQueryKey(nodeType)) ?? null;