When flagging, there are two special case options: spam and rude or abusive.
Note for moderators: some sites, including Stack Overflow, have more permissive, overriding guidance on the handling of these flags.
A post should be marked as spam only if it advertises a product, service, or similar and is unsolicited or lacks disclosure.
Due to the way search engines work, this includes links. For example, an otherwise normal post that contains a link to a website only in a punctuation mark is still spam.
Unsolicited means that mentioning the product serves no purpose other than promotion. For example, if an answer mentions a software that may solve the asker’s problem or a question about web programming references a website as an example, this is not unsolicited (it may still be spam if there is an undisclosed affiliation).
Lacks disclosure means that the author is clearly affiliated with the product but does not disclose their affiliation. Note that a simple “my” may suffice. However, the disclosure must happen in the post itself; the author’s username or profile do not count.
If an otherwise valid post contains an apparent spam link and especially if the bulk of the post is plagiarized from another post or from an off-site source, flag as spam (example). Do not try to salvage the post by removing the spammy content. If you're unsure, you can often find the original source with a Google search of its first sentence.
Be specifically cautious when judging posts falling into this category. Sometimes it may just be an innocent user leaving a signature trying to get some SEO. (Related)
It should not be marked as spam when:
The post contains no useful information, such as an answer that says “I don't care about your problem”. Flag as not an answer instead.
It contains only gibberish, such as “sdhsgfdhsfdshs”. Use the rude or abusive flag for these cases. See below.
A post should be marked as rude or abusive (formerly known as offensive) if it contains hate speech, obscenities, abuse against people, or abuse of the community or system, i.e., a clear violation of the be-nice policy.
Abuse of the system or community is everything that is created with the intention to harm them. This includes posts by new users that contain no useful content at all – i.e. gibberish posts along the lines of:
asdasdasdasdasdasdasdasdasdasdasdasd
As a rule of thumb, everything that would be out of place in polite discourse is rude or abusive.
If an otherwise valid post contains vulgar words as an expression of frustration, edit the bad part out instead of flagging the entire post as rude or abusive. If this results in an edit war or rollback war, flag for moderator attention.
Note that this is very different from handling an otherwise salvageable spam post.
Do not use this flag because:
A post criticizes somebody or something in a civil manner.
A post is a (civil) rant in disguise. If any part of the post can be salvaged, edit out the rant-y parts. If not, vote or flag to close as Opinion-based (for questions) or flag as not an answer (for answers).
Somebody appears to have posted nonsense due to an innocent mistake such as a copy-and-paste error.
The exact definitions of these terms are given above and the distinction can help moderators handle these flags if the problem is not blatantly obvious. (Note that if the spam or abuse is hidden, it may be necessary to elaborate this in a custom flag.)
Otherwise, the system does not differentiate between these flag types when counting the number of flags it has towards the thresholds to automatically downvote or delete. There are a couple unrelated cases where the system does differentiate between these flags:
Deleted answers with at least one helpful spam flag count towards automatic protection of a question. Answers with helpful rude or abusive flags, but no helpful spam flags, don't count.
If a post has a single validated "rude or abusive" flag, it won't be used as a review audit, to prevent NSFW posts from cropping up in the review queues.
Posts that are deleted by the system upon receiving six spam flags (or a single one from a moderator) may be used as review audits. As stated above, this doesn't apply for the "rude or abusive" flag.
Spam or rude or abusive flags (red flags) receive an extremely high priority in the moderation queue and come with severe penalties:
You can retract spam and rude or abusive flags like all other flags. They also expire after four days if the thresholds aren't reached. (Previously, they expired after two days - source.)
Rolling back a post to a previous state will revert to the number of flags from that particular revision. This allows the author (or someone else with edit rights) to rollback a post to which someone else introduced spam, rude, or abusive content in a later revision.
In addition, these flags can be cleared by moderators, whether active or already dismissed. This will cause the flags to be marked as disputed, even though they may have been marked helpful in the past. Since these flags carry heavy weight on the post and its author, a special mechanism is provided to clear borderline flags without penalizing anyone (declining the flag would penalize the flagger). In addition, this is sometimes used to remove bad audits from the review queues, as these flags can cause the posts to become review audits (while only posts that are actually deleted by the system in response to six spam flags are fed into the queues as known-spam audits, some spam is too subtle to expect reviewers to catch and can result in failed audits even if the reviewer is paying enough attention to ordinarily pass).
For reference:
At first this appears to be potentially a bad mechanic. An offensive edit would cause the OP to get punished with automatic rep loss. If the community isn't paying attention they could flag spam and cause unwarranted damage to the first revision owner.
There are ways to counter this. Before it reaches 6 flags, an editor could revert the edit and the flags would be revoked. However, if the 6 flags occur fast, or a mod flags, there's no chance to intervene.
The reasoning is in how edits can occur.
Any outliers to the above can be corrected and revoked by moderation.
Going with the above understanding, an editor is trusted. Likely they are trying to improve the post. So if they fail to completely improve the post and leave behind overlooked offensive material, they would incur the penalty in the suggested alternative system.
It is still likely for offensive material to get placed in an edit.
Considering the poor performance of reviewers quickly approving edits. It is possible that a malicious edit carefully hidden from first glance can be unintentionally approved. Resulting in the above problem.
If these suggestions are followed, we can avoid misdirected punishment and save correction efforts by moderation.
