Currently the shadertoy node does not run the updated ShaderToy GLSL code. In ShaderToy webpage- there are differences such as using fragCoord instead of gl_FragCoord etc. (https://community.vuo.org/t/-/5397 see @jstrecker 's response)
These differences are SUPER SMALL, yet without knowing how to fix even these super small issues a beginner would be lost. (I was lost when I first tried to get ShaderToy node to run ShaderToy.com code)
Instead of changing the ShaderToy node to comply with ShaderToy, (which potentially may change again) maybe ShaderToy can be repurposed into a generic GLSL image generator node, with similarities to the yet to be released GLSL node.
It is my understanding that the GLSL node in 1.3 will export a shader, whereas the re-purposed ShaderToy node (lets call it vuo.image.make.shader) would only export an image. I am always wanting to bring in 3D points into ShaderToy node. I am aware of the great solution that @jersmi came up with: https://community.vuo.org/t/-/5397 however, sometimes I just want to type a variable in and plug it in directly instead of cutting up text and formatting it.
My suggestion would be to bring the vuo.image.make.shader variables inline with GLSL node in 1.3, yet simply outputting an image.
Hopefully this would allow unlimited texture channels (currently shadertoy node is limited to 4) and custom variables such as in calculate maths node.
This feature would also allow a Fragment shader to be used easily either as texturing an objects surface, or as an image generator in its own right without re-coding variables, or unique ShaderToy GL input variables (between vuo.image.make.shader & vuo.shader.make).