Comment Guardian
Stop scam comments from reaching your customers.
Monitors comments on your Facebook Page for spam, phishing, scams, and impersonation. Auto-hides dangerous comments or alerts for review.
What is Comment Guardian?
Comment Guardian is a brand protection agent that monitors Facebook Page comments every 4 hours, classifies each as legitimate, suspicious, spam, phishing, or impersonation using AI, and auto-hides dangerous comments based on your configured rules. It supports three modes: monitor-only, approve-first, and auto-hide. Built for social media managers and brand safety teams.
10 min
Setup time
4 hours
Scan frequency
5
Classification categories
3
Operating modes
The problem — and how this kit solves it.
The Problem
Facebook Page comments are a magnet for spam, phishing links, and impersonators. A single "Contact us at +1-800-FAKE" comment on a popular post can redirect customers to scammers. Manual moderation does not scale — high-volume pages get hundreds of comments daily, and threats hide among legitimate engagement.
The Solution
This agent scans your Facebook Page comments every 4 hours and classifies each into 5 categories: legitimate, suspicious, spam, phishing, or impersonation. Three operating modes give you control: monitor-only (alerts without action), approve-first (flags for review), or auto-hide (automatically hides comments matching configured classifications — spam and phishing by default). A keyword watchlist adds custom detection rules. All actions are logged and sent via Slack alerts.
The Outcome
Remove spam and phishing comments from your Facebook Page within 4 hours of posting, before they reach your audience.
How this agent works
This agent runs autonomously on a schedule. Here's what happens each run.
Trigger
Each run includes
Scheduled comment scanning (every 4 hours by default)
AI classification: legitimate, suspicious, spam, phishing, impersonation
Three approval modes: monitor only, approve first, full auto
Auto-hide spam and phishing in full auto mode
Custom keyword watchlist for extra scrutiny
Slack alerts with comment previews and classification
Scheduled comment scanning (every 4 hours by default)
Completed in 2.1s
AI classification: legitimate, suspicious, spam, phishing, impersonation
Completed in 4.8s
Three approval modes: monitor only, approve first, full auto
Completed in 1.3s
Auto-hide spam and phishing in full auto mode
Completed in 5.4s
Custom keyword watchlist for extra scrutiny
Completed in 1.5s
Slack alerts with comment previews and classification
Completed in 1.3s
Here's what Comment Guardian actually produces.
Real output from a typical run — not a mockup.
Scanned 47 comments, flagged 5 (2 auto-hidden). 1 phishing ("Contact us at +1-800-FAKE"), 1 impersonation, 3 spam (crypto promotion).Three steps to autonomous.
Connect
Link your ad platform account via OAuth. Scoped permissions, encrypted credentials.
Configure
Set your targets, guardrails, and approval mode. Choose observe-only, approve-first, or full-auto.
Deploy
Activate the agent. It runs on schedule, analyzes data, and proposes or executes actions automatically.
Everything this kit does.
- Scheduled comment scanning (every 4 hours by default)
- AI classification: legitimate, suspicious, spam, phishing, impersonation
- Three approval modes: monitor only, approve first, full auto
- Auto-hide spam and phishing in full auto mode
- Custom keyword watchlist for extra scrutiny
- Slack alerts with comment previews and classification
Runs every 4 hours to scan recent comments
Who is this for?
Social media manager for a high-volume brand page
Auto-hide spam and phishing comments on posts that get hundreds of comments. The agent handles the volume while you focus on genuine community engagement.
Brand safety team at a regulated company
Use approve-first mode to flag suspicious comments for human review before action. Maintain compliance requirements while catching threats at the 4-hour interval.
Agency managing multiple client Pages
Deploy Comment Guardian per client Page. Each sends Slack alerts to the relevant channel. Auto-hide keeps client pages clean between your moderation shifts.
Frequently asked questions
Comments are classified as: legitimate (normal engagement), suspicious (potentially problematic but unclear), spam (promotional junk or bot content), phishing (links or contact info designed to steal credentials or redirect customers), or impersonation (someone pretending to be your brand or staff in comments).
Monitor-only sends alerts but takes no action on comments. Approve-first flags comments and waits for your approval before hiding. Auto-hide immediately hides comments matching your configured classifications (spam and phishing by default). You can switch modes at any time.
Yes. The default auto-hide classifications are spam and phishing. You can add impersonation and suspicious to the auto-hide list, or remove categories. The keyword watchlist lets you add custom terms that always trigger flagging regardless of AI classification.
A typical result: "Scanned 47 comments, flagged 5 (2 auto-hidden). 1 phishing ('Contact us at +1-800-FAKE'), 1 impersonation, 3 spam." Each flagged comment shows the original text, classification, and action taken.
No. Comment Guardian monitors Facebook Page comments only. It requires a Facebook Page ID during setup and uses the Meta Graph API for comment access.
Pairs well with.
These kits complement each other. Deploy them together for coverage no single agent can provide.
Brand Impersonation Monitor
Comment Guardian protects your real pages; Impersonation Monitor catches fake pages pretending to be you
Learn moreMeta General Campaign Optimizer
Protect your Page engagement quality while optimizing your paid campaign performance
Learn moreYou might also want.
Brand Impersonation Monitor
Searches Facebook for pages using your brand name, then uses AI to classify each as critical/high/medium/low risk. Official pages from connected Facebook Page integrations are automatically whitelisted. False positives can be marked as official from the run history.
Facebook Ads Campaign Health Check
This is the smallest viable Facebook Ads action agent for Zimmer. Instead of asking for raw Graph URLs or a pasted token, it uses your existing Meta Ads integration. You choose one synced campaign, the agent fetches the latest campaign details and insights through an internal integration-backed health-check endpoint, computes budget utilisation, and writes a structured summary into the run log.
Meta General Campaign Optimizer
The General Campaign Optimizer is the first write-owner optimizer in Zimmer. It reads your full campaign state — budgets, spend, KPIs, pacing — and proposes concrete actions: pause underperformers, boost top performers, rebalance budget across ad sets. It enforces one optimizer per campaign to prevent conflicting actions, and every proposal passes through validation, guardrails, and your chosen approval mode.
Ready to deploy?
Comment Guardian is one of 100+ starter kits.
Describe what you need — or deploy this one now. Every kit runs with guardrails, approval workflows, and full audit logging.
Browse all starter kits