Right now I find myself forgetting available commands and having to manually look through my .lute folder (along with directory and sometimes test file clutter) is tedious. I would love something built into the CLI to handle this for me.
As an example of what I'm looking for, here's the contents of Flipbook's .lute directory:
.lute
├── analyze.luau
├── build.luau
├── bump-version.luau
├── clean.luau
├── install.luau
├── lib
├── lint.luau
├── serve-docs.luau
├── tasks
└── test.luau
Then when running any of the following:
lute run --list # available scripts
lute run --help # help text _and_ available scripts
lute run # same as the last
I would expect to get some output like:
Available commands:
- analyze
- build
- bump-version
- clear
- install
- lint
- serve-docs
- test
Run any of these with `lute run <scriptName>`
Anecdotally I also see my agents get confused and start running commands like lute run .lute/build.luau and, while this executes just fine, its clear the agent doesn't fully understand how lute run works and that it's nicer for a human to see lute run build fly by than a file path. Having output like the above would help steer them (and myself) in the right direction.
Right now I find myself forgetting available commands and having to manually look through my
.lutefolder (along with directory and sometimes test file clutter) is tedious. I would love something built into the CLI to handle this for me.As an example of what I'm looking for, here's the contents of Flipbook's
.lutedirectory:Then when running any of the following:
I would expect to get some output like:
Anecdotally I also see my agents get confused and start running commands like
lute run .lute/build.luauand, while this executes just fine, its clear the agent doesn't fully understand howlute runworks and that it's nicer for a human to seelute run buildfly by than a file path. Having output like the above would help steer them (and myself) in the right direction.