Loona AI Comment Manager—Replies, Spam Shield & Owner Alerts

    1

    The Loona AI Comment Manager is the workflow that handles all three — calmly, in your brand voice, on your Facebook Page, Instagram account or YouTube channel.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Loona AI Comment Manager—Replies, Spam Shield & Owner Alerts

    Example session with this skill installed

    Check the new comments on my Facebook Page and Instagram. Use the default friendly voice. Note that we don't have pricing for the custom cakes yet.

    • Read your context and instructions
    • Compiled the loona ai comment

    I have processed 5 new comments across Facebook and Instagram. No incidents detected.

    Run Results

    COMMENT: "Your cupcakes are the best in the city! 🧁" (under: Strawberry Launch Post)
    CLASS: Praise — genuine compliment about product quality.
    REPLY DRAFT: "Thank you so much! We are thrilled you enjoyed them. Which flavor did you try first?"
    ACTION: None

    COMMENT: "How much for a 3-tier custom wedding cake?" (under: Wedding Portfolio)
    CLASS: Question — specific product inquiry.
    REPLY DRAFT: "Hi! Thanks for reaching out about our custom wedding cakes. I will confirm the current pricing for a 3-tier cake for you — what is the date of your event?"
    ACTION: None

    COMMENT: "Get cheap followers at link-scam.com!!!" (under: Strawberry Launch Post)
    CLASS: Spam/Profanity — external link/unsolicited promotion.
    ACTION: FLAG FOR HIDING — link spam

    COMMENT: "My order #1234 was totally wrong and I want a full refund now." (under: General Post)
    CLASS: Complaint — specific order issue.
    REPLY DRAFT: "We are very sorry to hear about the trouble with your order. The owner will review this personally and get back to you via direct message to resolve it."
    ACTION: None

    COMMENT: "I'm going to sue you because your floor was wet." (under: General Post)
    CLASS: Owner-Decision — legal threat and safety claim.
    ACTION: ESCALATED TO OWNER — safety/legal claim; evidence log #1

    OWNER SUMMARY
    Incident: none
    Checked: 5 new comments (last 24 hours)
    Replies drafted: 3 (1 complaint draft below)
    Flagged for hiding: 1 — link spam
    Escalated to owner: 1 — evidence log #1
    Placeholders to fill: [custom cake pricing]
    Defaults used: Brand voice, Hide policy, Reply policy

    EVIDENCE LOG
    #1 — Facebook — General Post
    Commenter: John Doe
    Date/time: 2023-10-24 09:15 AM
    Escalation reason: Legal threat/safety claim
    Verbatim: "I'm going to sue you because your floor was wet."

    Next steps

    1. Review and post the 2 routine drafts for Praise and Questions.
    2. Review the Complaint draft regarding order #1234 before sending.
    3. Address the legal escalation in evidence log #1 immediately.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Your comment section is your shop window. Every genuine comment deserves a warm reply, every spam comment makes the page look abandoned, and one wrong reply to a sensitive comment can cost you money, a dispute, or your reputation. The Loona AI Comment Manager is the workflow that handles all three — calmly, in your brand voice, on your Facebook Page, Instagram account or YouTube channel.

    On every check it collects the new comments and sorts each one into exactly one class: genuine praise, a genuine question, a complaint, spam or profanity, or an owner-decision matter. Praise and questions get short, friendly draft replies in the commenter's own language and script — English, Spanish, Arabic, French, German and Portuguese templates are included — answered only from facts you provided, never invented. Complaints get an empathy-first draft held separately for your review, written under strict guardrails: it never admits fault or causation in writing, never names or blames a staff member, and never calls a commenter a liar. Spam and profanity are flagged for hiding under your own policy — flagged only, or hidden automatically, as you set it.

    One rule is absolute: owner-decision comments are never replied to, never drafted for, and never hidden. Refund demands, disputes, business offers, health, safety and discrimination claims reach you verbatim and untouched in your owner summary, each with an evidence log.

    And on a bad day, the system does not improvise. Incident Mode switches the run by severity: a spike is watched, a wave of near-identical negatives is handled as one cluster with a single representative draft, and a crisis — extortion, coordinated attacks, safety threats — stops all drafting and sends you an alert with the evidence preserved. Extortion messages are marked and preserved, never answered, and paying is never suggested.

    Every run ends with a short owner summary: what was checked, what was drafted, what was flagged, what was escalated — so five minutes a day keeps your comment section clean, answered and safe. Built by Loona Automation from its live daily comment operations. Version 1.1.1.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Social Media.

    No DM support. Short sarcasm may be misclassified. Requires owner-provided baseline for incident triggers. Does not provide legal or PR advice.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed today

    Permissions

    No special permissions declared or detected

    Listedtoday

    What's inside

    Frequently Asked Questions

    The skill integrates with any agent platform connected to Facebook, Instagram, or YouTube APIs, and is designed to work with all major LLM models that support structured classification tasks.

    Incident Mode is a specialized triage layer that activates during high-volume negativity or review-bombing, grouping similar comments into clusters and preserving evidence for the owner rather than attempting individual automated replies.

    No, the skill is designed with a strict 'placeholder rule' that prevents it from answering legal threats, health claims, or financial demands, ensuring these sensitive issues are escalated to you verbatim.

    Yes, the skill automatically detects the commenter's language and script, drafting a reply in that same language while maintaining your brand's specific tone and voice.

    You will receive the complete logic framework for the decision matrix, the Incident Mode escalation protocols, and the structured summary templates needed to deploy the skill to your agent.

    While the skill is highly accurate, it may struggle with very short, sarcastic comments; in these ambiguous cases, it is programmed to flag the comment for your manual review rather than risk a misaligned reply.

    New in Social Media