Two things are true at once: platforms genuinely restrict tags, and almost every "banned hashtag list" you will find online is years out of date and partly invented. Both facts matter.
What restriction actually looks like
Platforms limit tags that attract policy-violating content. The restriction takes several forms — the tag page shows only top posts with no recent section, the page shows a content warning, the page shows nothing, or the tag simply stops surfacing posts in search.
The consequence for you is that a post carrying a restricted tag can have its distribution quietly reduced. There is no notification. This is why an unexplained drop sometimes traces back to one tag.
Why the lists do not work
Restriction status changes constantly and is not published. Tags get restricted, appeals succeed, new ones get added, and lists compiled in 2021 describe a state that no longer exists. Worse, most circulating lists are copied from each other without verification.
The check that does work
Search the tag on the platform, from the app, before you use it. Thirty seconds. You are looking for: a recent-posts section that exists and is populated, content that resembles yours, and no warning message. If any of those fail, drop the tag.
Do this for every unfamiliar tag you plan to use repeatedly. You do not need to do it for tags you have been using successfully for months.
The surprising ones
Restricted tags are frequently innocuous words that got adopted by communities posting prohibited content. Ordinary vocabulary around bodies, health, wellness and dating is disproportionately represented. The tag being harmless in your usage does not change its status.
If you think you have been affected
Remove the suspect tag, wait for the next posting cycle, and compare. Editing tags on an existing post rarely restores its distribution — the initial classification has already happened. The value of the check is in the next post, not the current one.
What this generator does
It builds tags from ordinary vocabulary banks and does not draw from known-restricted terminology. That reduces the risk substantially but cannot eliminate it, because platform lists are private and change without notice. Verify anything you plan to use heavily.