Handle Trolls, Spam and Harassment in Comments with Ready Replies

You get a decision table that separates criticism, trolling, spam and harassment, plus ready replies, a no-reply rule for each type, and clear report, block and document steps.

Prompt

Prompt
You are a community manager who moderates comments for creators and small brands.

Help me build a moderation playbook for my comments. Before writing anything, ask me these questions in one message and wait for my answers:
1. Which platforms, and what kind of account (creator, small business, brand)?
2. What tone should replies have (warm, dry, formal)?
3. Paste 3-6 real problem comments I've received, or describe them.
4. Do I have written community rules?

Rules:
- Sort every comment into one of four types: fair criticism, trolling, spam, harassment. Give a one-line test that tells each type apart.
- Fair criticism always gets a public reply. Trolling gets at most one reply. Spam and harassment get no public reply.
- Every reply is under 40 words, has no sarcasm, and states facts once without arguing.
- For each type, state the action: hide, delete, report, block, or document (screenshot with date, username and link).
- Say when to escalate to platform support or police: threats, doxxing, stalking, or coordinated pile-ons.

Output format:
1. Decision table: Type | How to spot it | Reply or no reply | Report / Block / Document
2. Ready replies: 2 per type that gets a reply
3. My comments sorted: Comment | Type | Action | Reply or "no reply"
4. Escalation checklist (5 items)

End with one line naming gray areas I should judge myself, and note this is not legal advice; contact police or a lawyer for threats.

Example output

Example for: a small bakery on Instagram and Facebook, warm tone, no written community rules.

1. Decision table

TypeHow to spot itReply or no replyReport / Block / Document
Fair criticismNames a specific problem you could fixReply publicly within 24 hoursNone
TrollingWants a reaction, not a fix; vague insults or baitOne calm reply at most, then stopHide if repeated; block after 3
SpamLinks, "DM me," fake giveaways, unrelated promosNo replyDelete, report as spam, block
HarassmentTargets a person, slurs, threats, personal detailsNo replyScreenshot, report, block; escalate if threats

2. Ready replies

Fair criticism

  • "Sorry the croissants were stale on Saturday. That's not our standard. Please DM us your order time so we can make it right."
  • "Thanks for flagging the wait. We've added a second person at the counter on weekends."

Trolling

  • "We're happy with our recipes, but thanks for stopping by."
  • "Fair enough, not every bakery suits everyone. Have a good day."

3. Your comments sorted

CommentTypeActionReply
"$9 for a cake slice?? Rip-off"Fair criticism (price)Reply once"We use local butter and bake daily, which shows in the price. Our day-old shelf is half off after 3pm."
"Bet this is all frozen dough lol"TrollingOne reply, then ignore"Everything's made from scratch at 5am. You're welcome to watch through the window."
"Earn $500/day from home, link in bio"SpamDelete, report, blockNo reply
"Tell Maria I know where she parks"Harassment (threat)Screenshot, report, block, escalateNo reply
"Worst service, never coming back"Fair criticism (vague)Reply and invite details"Sorry to hear that. Could you DM us what happened so we can look into it?"

4. Escalation checklist

  1. Screenshot the comment, username, profile URL, date and time before deleting.
  2. Save everything in one dated folder.
  3. Report through the platform's harassment or threat option and note the report ID.
  4. Tell any named staff member and agree on safety steps such as parking and closing shifts.
  5. For threats of violence or stalking, contact local police with your folder.

Gray areas: sarcastic regulars and heated price complaints need your judgment. This is not legal advice; contact police or a lawyer about threats.

Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • The four-type decision table with one-line tests stops you from feeding trolls or ignoring real customers.
  • Hard reply rules (under 40 words, one reply max to trolls, no sarcasm) keep responses calm and short.
  • Sorting your own pasted comments turns the playbook into decisions you can act on today.

Use this in ChatGPT, Claude or Gemini when a post draws hostile or junk comments. Have a few real comments copied and know your platform and preferred reply tone.

Common edits

  • Add a fifth type for misinformation with a reply that links to a source.
  • Ask for a version formatted as a one-page guide for team moderators.

Prompt details

By
Prompts-GPT editors
Published

Did this prompt work?

Keep exploring

More in Social Media