I was playing around with Bing's Image Creator and was able to come up with the photos attached.

The problem is I keep getting the dreaded, "prompt blocked" warnings, knowing that once I reach a certain number, my account will be permanently blocked.

It's so strange, as it will accept 'people playing in mud' or 'men playing in mud' but as soon as I put 'women playing in mud' it has a problem. Same with other words. I can have a woman in a tub of oatmeal, but if I change oatmeal to chocolate, it suddenly blocks me.

Any work-around for this? I'm the newest of newbies at the moment.

Not terrible for a first attempt. Bing Image creator and Dalle through ChatGPT have different content filters applied, that said some of the more common filters are shared. pictures of woman are heavily weighted towards nsfw as soon as you add other prompts. Over the last week, their content filter has become more an more aggressive with some key words being black listed.
Slime was allowed last week but quickly blocked after 4chan degens started creating tenticle porn. Cream pie for obvious reasons too.
Drenched doesn't work for me, but soaked or engulfed does.

You can go down the route of using image in the style of 'X' artist and add other sfw criteria to help steer the prompt back on course.
If you are going for more photo realistic, then I'm finding an iterative approach works quite well. Specifying specific clothes can be hit an miss but prompting for a specific style/look like preppy 1990's or emo 2000's and then adding specfic clothes works well.
For mess, I start iteratively adding items, using positive sfw adjectives and adverbs key words helps steer the model back on course. This like playfully, having fun, smiling, laughing.
Location plays a big part too I've found, a parisian park in summer time or a cozy cafe in the evening works well but bedrooms not so good.
Its all trial and error though. There are 2 layers of content filtering with dalle. They filter the original prompt and then the images generated are also filtered by a secondary AI. With ChatGPT you can refine the prompt and ask it to add additional tweaks and changes to the image and regenerate. Earlier last week, you could also see the prompt feedback and gauge where the issues are coming from.

It's an interesting time though, the power is undoubtedly there in the underlying model and unfiltered would be incredible but you can understand why they are struggling to moderate it if you look on 4chan and other sites. For now, they're being over protective. It's still early days though and we could see the content filter swing both ways.

you just have to put up with it ive had cake batter be okay but chocolate cake batter blocked