ci: auto-minimize activity-farming spam comments (#7560)

A 26-account network posted 84 emoji-padded template comments across 20
issues over ten hours, burying real reports. The accounts are structurally
distinct from real reporters: recently registered, zero followers, a pile of
forked repos, and sub-50 character comments with no code, link, or log.

Replayed against all 7604 issue comments in the repo, requiring every signal
to fire catches all 84 spam posts with no false positive. Any single signal
relaxed pulls in real users.
This commit is contained in:
YHH authored and GitHub committed 2026-08-05 12:24:55 +08:00
1 parent cb4a21306a
commit 2206ee654f
1 file changed
+64
+64
View File
@@ -0,0 +1,64 @@
name: Spam comment guard
# Activity-farming bots post short emoji-padded template comments on whatever
# issue is currently active. They are cheap to spot structurally: a recently
# registered account with no history here, no followers, and a pile of forked
# repos, posting a sub-50 character comment with emoji and no code/link/log.
# Every signal must fire — on 7604 replayed comments that caught all 83 known
# spam posts with no false positive. Loosening any one of them reintroduces
# real users ("点赞", "同样的问题", a 2020 account posting "感谢各位❤").
on:
issue_comment:
types: [created]
permissions:
issues: write
pull-requests: write
concurrency:
group: spam-guard-${{ github.event.comment.id }}
jobs:
screen:
if: github.event.comment.user.type != 'Bot' && github.event.comment.author_association == 'NONE'
runs-on: ubuntu-latest
steps:
- uses: actions/github-script@v9
with:
script: |
const EMOJI = /[\u{1F300}-\u{1FAFF}\u{2600}-\u{27BF}\u{2B00}-\u{2BFF}\u{FE0F}]/gu;
const comment = context.payload.comment;
const body = comment.body || '';
const hasSubstance = /```|https?:\/\/|!\[|^>/m.test(body);
const emojiCount = (body.match(EMOJI) || []).length;
const bare = body.replace(EMOJI, '').trim();
const { data: author } = await github.rest.users.getByUsername({
username: comment.user.login,
});
const ageDays = (Date.now() - Date.parse(author.created_at)) / 86400000;
const signals = {
newAccount: ageDays < 365,
noFollowers: author.followers === 0,
forkPadded: author.public_repos >= 15,
tooShort: bare.length <= 45,
emojiPadded: emojiCount >= 1,
noSubstance: !hasSubstance,
};
const failed = Object.entries(signals).filter(([, v]) => !v).map(([k]) => k);
core.info(`${comment.user.login}: ${failed.length ? `kept (${failed.join(',')})` : 'spam'}`);
if (failed.length) return;
await github.graphql(
`mutation($id: ID!) {
minimizeComment(input: { subjectId: $id, classifier: SPAM }) {
minimizedComment { isMinimized }
}
}`,
{ id: comment.node_id },
);
core.notice(`Minimized spam comment by ${comment.user.login}`);