Reading the pre-flight risk check
Where the score comes from, why every line names its source, and what to do with a high one.
The score is a probability, not a verdict
The number from 0 to 100 estimates how likely this post is to be removed. It is not a prediction that it will be. Automated moderation settings are private, so nothing can honestly promise otherwise. Treat a high score as a reason to read the listed reasons.
Every line says where it came from
Each finding is tagged with its source: the community's own rules, your account, your posting history here, the text itself, the community's AI policy, or what other users have measured. A rule-based finding is a fact. A crowd-based one is a sample.
Blockers, warnings and notes
A red cross is something the published rules forbid outright, such as a link in a text-only community. An amber mark is a risk worth weighing. Blue notes are context. Fix the crosses first: they are the cheapest to fix and the most likely to matter.
What “not confirmed” means here
When the extension cannot read something, because of rate limits, a private wiki page or a community it has never seen, it lists the gap instead of guessing. Missing information is never rendered as a clean bill of health.
The optional AI rule review
Once the title is written and the community's rules are loaded, you can have the AI read your draft against each rule and your account, for 1 credit. It catches what a keyword scan cannot: an off-topic post, low effort, promotion dressed up as a question. Each finding names the rule it refers to and joins the same score with a lower weight than the published rules. Edit the draft and the review stops counting until you run it again.