Evaluate Tool - Nushell

Execute a command in the nushell. Successful evaluations return a structured NUON record string with metadata fields including cwd, history_index, timestamp, and either output or note. When evaluation fails, the MCP tool response is marked as an error and the content contains a structured NUON error record.

Avoid commands that produce a large amount of output, and consider piping those outputs to files. If you need to run a long lived command, background it - e.g. job spawn { uvicorn main:app } so that this tool does not run indefinitely.

Command Equivalents to Bash:

Bash Command	Nushell Command	Description
`mkdir -p <path>`	`mkdir <path>`	Creates the given path, creating parents as necessary
`> <path>`	`o> <path>`	Save command output to a file
`>> <path>`	`o>> <path>`	Append command output to a file
`> /dev/null`	`ignore`	Discard command output
`> /dev/null 2>&1`	`o+e>\| ignore`	Discard command output, including stderr
`command 2>&1`	`command o+e>\| ...`	Redirect stderr to stdout (use `o+e>` or `out+err>`)
`cmd1 \| tee log.txt \| cmd2`	`cmd1 \| tee { save log.txt } \| cmd2`	Tee command output to a log file
`command \| head -5`	`command \| first 5`	Limit the output to the first 5 rows of an internal command (see also last and skip)
`cat <path>`	`open --raw <path>`	Display the contents of the given file
`cat <(<command1>) <(<command2>)`	`[(command1), (command2)] \| str join`	Concatenate the outputs of command1 and command2
`cat <path> <(<command>)`	`[(open --raw <path>), (command)] \| str join`	Concatenate the contents of the given file and output of command
`for f in *.md; do echo $f; done`	`ls *.md \| each { $in.name }`	Iterate over a list and return results
`for i in $(seq 1 10); do echo $i; done`	`for i in 1..10 { print $i }`	Iterate over a list and run a command on results
`cp <source> <dest>`	`cp <source> <dest>`	Copy file to new location
`rm -rf <path>`	`rm -r <path>`	Recursively removes the given path
`date -d <date>`	`"<date>" \| into datetime -f <format>`	Parse a date (format documentation)
`sed`	`str replace`	Find and replace a pattern in a string
`grep <pattern>`	`where $it =~ <substring>` or `find <substring>`	Filter strings that contain the substring
`command1 && command2`	`command1; command2`	Run a command, and if it's successful run a second
`stat $(which git)`	`stat ...(which git).path`	Use command output as argument for other command
`echo /tmp/$RANDOM`	`$"/tmp/(random int)"`	Use command output in a string
`cargo b --jobs=$(nproc)`	`cargo b $"--jobs=(sys cpu \| length)"`	Use command output in an option
`echo $PATH`	`$env.PATH (Non-Windows) or $env.Path (Windows)`	See the current path
`echo $?`	`$env.LAST_EXIT_CODE`	See the exit status of the last executed command
`export`	`$env`	List the current environment variables
`FOO=BAR ./bin`	`FOO=BAR ./bin`	Update environment for a command
`echo $FOO`	`$env.FOO`	Use environment variables
`echo ${FOO:-fallback}`	`$env.FOO? \| default "ABC"`	Use a fallback in place of an unset variable
`type FOO`	`which FOO`	Display information about a command (builtin, alias, or executable)
`\`	`( <command> )`	A command can span multiple lines when wrapped with ( and )

If the polars commands are available, prefer it for working with parquet, jsonl, ndjson, csv files, and avro files. It is much more efficient than the other Nushell commands or other non-nushell commands. It exposes much of the functionality of the polars dataframe library. Start the pipeline with plugin use polars

An example of converting a nushell table output to a polars dataframe:

ps | polars into-df | polars collect

An example of converting a polars dataframe back to a nushell table in order to run other nushell commands:

polars open file.parquet | polars into-nu

An example of opening a parquet file, selecting columns, and saving to a new parquet file:

polars open file.parquet | polars select name status | polars save file2

Important: The glob command should be used exclusively when you need to locate a file or a code reference, other solutions may produce too large output because of hidden files! For example do not use find or ls -r. Use command_help tool to learn more about the glob command.

Important: Variables and environment changes persist across tool calls (REPL-style). You can set a variable in one call and access it in subsequent calls:

let x = 42 in one call, then $x in the next call returns 42
$env.MY_VAR = "hello" persists for later calls

However, external processes run in their own environment. Use absolute paths when possible.