One thing that I do frequently is run a command that produces a lot of output, often curl or nmap, which I would rather not run again. This means that instead or re-running the command and redirecting the output to the kill ring or clipboard, I have to carefully select the output from the buffer and then M-w (kill-ring-save) it. That's more effort than I'd like, so I made a function to select the previous command's output and add it to the kill ring:

(defun eshell/kill-previous-output (&optional nth)
  "Copies the output of the previous command to the kill ring.
When nth is set, it will copy the nth previous command."
    ;; Move to the end of the eshell buffer.
    (goto-char (point-max))
    ;; Move to the start of the last prompt.
    (search-backward-regexp eshell-prompt-regexp nil nil nth)
    ;; Move to the start of the line, before the prompt.
    ;; Remember this position as the end of the region.
    (let ((end (point)))
      ;; Move to the start of the last prompt.
      (search-backward-regexp eshell-prompt-regexp)
      ;; Move one line below the prompt, where the output begins.
      ;; Find first line that's not blank.
      (while (looking-at "^[[:space:]]*$")
      ;; Copy region to kill ring.
      (copy-region-as-kill (point) end)
      ;; Output stats on what was copied as a sanity check.
      (format "Copied %s words to kill ring." (count-words-region (point) end)))))

This function doesn't use Eshell's builtin functions, because they break in the common case of the previous command containing a newline. Here's why:

(defun eshell-next-prompt (n)
  "Move to end of Nth next prompt in the buffer."
  (interactive "p")
  (forward-paragraph n)

As you can see, the call to forward-paragraph messes things up for my use case.