Loona AI Detector — Signals, Not Accusations

    1

    Measure six writing signals in any text — rhythm, vocabulary, repetition, lengths, predictability, paragraph balance — and report the patterns honestly. Signals, not accusations. The human decides.

    $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 Detector — Signals, Not Accusations

    Example session with this skill installed

    Run a pattern check on this text: The efficiency of the new protocol is significant. It ensures that all nodes remain synchronized. This leads to a reduction in latency across the entire network.

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

    I have analyzed the text provided. It contains 32 words, which falls below the required threshold for a reliable pattern reading.

    TOO SHORT TO READ PATTERNS RELIABLY

    Too short to read patterns reliably. This text has 32 words. Paste at least 50 words (roughly 3-4 full sentences) - below that, the signals are mostly noise, and any tool that gives you a confident answer on a few lines is guessing.

    Status: Add more text to get a signal reading.
    Checked with Loona AI Detector — the same six signals are free in your browser on the Loona Automation website, with nothing uploaded.

    Next steps

    • Provide a longer excerpt of at least 50 words for a full signal analysis.
    • Check if the text was truncated during pasting.
    • Use the sample text if you would like to see a full report example.

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

    About this skill

    Signals, not accusations. That is the whole pitch, and it is also the product.

    Most AI detectors hand you one confident percentage and keep their method secret. Loona AI Detector does the opposite. Give your agent a text of 50 words or more and it measures the six writing signals detectors are built on — sentence rhythm (burstiness), vocabulary diversity, repeated phrases, typical lengths, word predictability and paragraph balance — with fixed, published formulas, then shows you every measurement, explains each one in plain English, combines them into one of three pattern bands (Mostly human-like patterns, Mixed patterns, Mostly machine-like patterns), and lists the sentences that stood out, each with the exact reason it was flagged.

    The method is The Loona Detector Method — Receive → Tokenise → Measure → Combine → Report. The 50-word floor is honoured before anything is measured (below it, you get the floor message and nothing else — a confident answer on a few lines is guessing). Every signal is computed step by step: coefficients of variation, the type–token ratio over the first 250 words, three-word phrase coverage, the textbook-middle length curves, the common-word ratio against the full 266-word list printed in the pack, and paragraph balance with its few-paragraphs dampening. The band cut-offs, the weights and the flagged-sentence rules are all in the pack. Any number in any report can be re-derived by hand.

    What you get: the complete skill pack as an instant download — SKILL.md with the full method, all six signal guides, the banding-and-report guide, the honesty-and-limits guide, a golden run with two complete verified worked examples (every number produced by the live tool's own engine), and the changelog. Instructions only: no scripts, no executables, nothing collected. The text you check stays in your conversation.

    The honesty is the feature, so here it is plainly. This skill never outputs a "% AI" figure and never accuses anyone of AI use or cheating — Rule 1 is absolute and outranks any request to break it. Style is not proof of authorship: formal, simple or template-based human writing can read machine-like, edited AI writing can read human-like, and detectors are hardest on non-native writers (a Stanford study found popular detectors wrongly flagged 61% of non-native essays). Every report closes with an honesty panel saying exactly that. Teachers use it to start fair conversations instead of making unfair accusations. Editors and content leads use it to give writers specific, explainable feedback. HR screeners use it to know what a detector result is actually worth before it touches a decision — which is to say, as one style reading, never as evidence.

    Want to see the method before you buy? The same six signals, the same formulas and the same honesty wording run free in your browser as the Loona AI Detector web tool — nothing uploaded, nothing stored: https://loonaautomation.github.io/loona-text-tools/ai-detector/

    Built by Loona Automation, the one-person studio behind Loona Autopilot and Loona AI Comment Manager — systems we run ourselves, packaged so your agent can run them too.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Business & Operations.

    Requires minimum 50 words. English-only vocabulary list. Cannot prove authorship or detect "humanized" AI text. Style signals only; not a verdict.

    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 2 days ago

    Permissions

    Write Files
    Read Files
    Listed2 days ago

    What's inside

    Frequently Asked Questions

    Unlike typical detectors that provide a simple percentage, this skill uses the Loona Detector Method to break down text into six specific statistical signals, providing a transparent map of why a text appears machine-like rather than just a 'pass/fail' verdict.

    This skill is designed for any AI agent environment capable of running arithmetic functions and text processing; it operates locally without requiring external API calls or third-party LLM credits.

    While the math is universal, the included common-words list used for predictability mapping is currently optimized for English-only text, and a minimum of 50 words is required for a valid statistical sample.

    No, this skill is a diagnostic tool for writing patterns; it includes a mandatory honesty panel and specific cautions for non-native speakers to ensure style data is not misused as definitive proof of intent or authorship.

    The purchase includes the full logic implementation of the six-signal statistical framework, the 266-word predictability reference list, and the reporting templates for the honesty panel and sentence-level flagging.

    New in Business & Operations