fix(mcp): drop the success-path status reconciliation heuristic

The client-side serversList invalidation on a successful probe assumed the probe
updated the stored status, but a server-side cache hit returns tools without
touching status — so in the failed-cache-delete edge it fired a pointless
refetch. The benefit (instant vs the 60s serversList stale-time for status
recovery) doesn't justify the incorrect assumption; rely on the existing 60s
stale-time + SSE push for status recovery instead. Keeps the corrected keep/drop
comment.

Author: waleedlatif1

apps/sim/hooks/queries/mcp.test.tsx

@@ -176,34 +176,6 @@ describe('useMcpToolsQuery', () => {
     unmount()
   })
 
-  it('refreshes the server list when a previously-failed server discovers successfully', async () => {
-    let listCalls = 0
-    mockRequestJson.mockImplementation(async (contract) => {
-      if (contract === listMcpServersContract) {
-        listCalls++
-        return {
-          success: true,
-          data: {
-            servers: [server('s1', { authType: 'headers', connectionStatus: 'disconnected' })],
-          },
-        }
-      }
-      if (contract === discoverMcpToolsContract) {
-        return { success: true, data: { tools: [{ name: 'tool-a', serverId: 's1' }] } }
-      }
-      throw new Error('Unexpected MCP request')
-    })
-
-    const { unmount } = renderHookWithClient(() => useMcpToolsQuery(WORKSPACE_ID))
-    await flush()
-
-    // A successful probe against a server the cache still shows 'disconnected' refreshes the
-    // list (server-side the status is now 'connected') so the row clears its red state.
-    expect(listCalls).toBeGreaterThanOrEqual(2)
-
-    unmount()
-  })
-
   it('refreshes the server list after a connected OAuth discovery fails', async () => {
     let serverListRequests = 0
     mockRequestJson.mockImplementation(async (contract) => {

apps/sim/hooks/queries/mcp.ts

@@ -172,19 +172,7 @@ export function useMcpToolsQuery(workspaceId: string) {
       queryKey: mcpKeys.serverToolsList(workspaceId, serverId),
       queryFn: async ({ signal }: { signal?: AbortSignal }) => {
         try {
-          const tools = await fetchMcpTools(workspaceId, false, signal, serverId)
-          // A successful probe flips the stored status to `connected` server-side; if the
-          // cached list still shows this server failed, refresh it so the row clears its red
-          // state (and its tools stop being dropped) instead of waiting out the list stale-time.
-          const cached = queryClient.getQueryData<McpServer[]>(mcpKeys.serversList(workspaceId))
-          const status = cached?.find((s) => s.id === serverId)?.connectionStatus
-          if (status && status !== 'connected') {
-            queryClient.invalidateQueries(
-              { queryKey: mcpKeys.serversList(workspaceId) },
-              { cancelRefetch: false }
-            )
-          }
-          return tools
+          return await fetchMcpTools(workspaceId, false, signal, serverId)
         } catch (error) {
           await queryClient.invalidateQueries(
             { queryKey: mcpKeys.serversList(workspaceId) },