Usage

Start the shell:

yupsh

Or drive it non-interactively (pipes and redirects use plain line scanning, so scripts are deterministic):

yupsh < script.txt

Input model

A pipeline’s input is chosen by its first stage:

First stage	Input source
A source command (`echo`, `seq`, `ls`, `find`, `yes`, `emit`)	the command generates the stream
A filter with file arguments (`cat file.txt`, `wc -l a.txt`)	the named files, opened via the framework’s `ByteFileSource`
A filter with no file arguments (`grep foo`)	standard input

Later stages must be filters — a source after a pipe, or file arguments on a non-first stage, are reported as errors.

Shell expansion

Globbing: unquoted *, ?, and […] expand against the working directory. A pattern with no matches is left literal (POSIX default).
Tilde: a leading ~ or ~/… expands to the home directory.
Quoting: single or double quotes group arguments and suppress expansion (echo '*.go' prints *.go).

yup> ls *.go
yup> wc -l *.md
yup> cat ~/notes.txt | grep TODO

Examples

# sources
yup> echo hello world
yup> seq 1 10

# pipelines
yup> seq 1 10 | grep -v 5 | head -n 3
yup> seq 1 100 | wc -l
yup> echo HELLO | tr A-Z a-z

# files and globs
yup> wc -l *.go
yup> cat data.txt | sort | uniq

# subprocess escape hatches
yup> echo hi | exec cat
yup> echo hello | perl -p 's/l/L/g'

Built-ins

help — list commands and usage
version — show version
clear — clear the screen
exit / quit — leave the shell
lines beginning with # are ignored