[qodelicious]

split column names its columns from column0 now (nushell/nushell#17064), so a few modules were still reading the wrong column.

• from cpuinfo replaced spaces in the values instead of the keys, so cpu MHz stayed as is and flags turned into one long fpu_vme_de... string. It also used get column1 / get flags inside update closures, which now get the cell value and not the row, so it errored on any input. Switched to str replace ... column0 and update flags { split row ' ' }.
• from dmidecode lowercased the values instead of the keys
• clone-all asked for column2 of org/repo, which doesn't exist anymore. grab repo name was also typed as string -> string but is called with no input, so clone all failed with "Pipeline empty". Changed it to nothing -> string.
• fuzzy-command-search took the description instead of the command name from the fzf line
• the full-line prompt read the path column instead of the XY status field from git status --porcelain=2, so the staged and modified counts were off
• AutoASPM took column2 of the lspci -xxx lines, the hex bytes are in column1 now

I tested each one with nu 0.116.1 against sample input (a fake /proc/cpuinfo, dmidecode output, porcelain v2 lines, lspci lines and a stubbed gh). They all error or give wrong output on main and give the right result with this change. toolkit check pr passes.

[fdncred]

Thanks