More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    LinkedIn Humanizer: Remove AI Tells

    1

    Score and strip the patterns that make a LinkedIn draft read AI-written — density vocabulary, reveal bridges, fragment stacks, performed sincerity — without flattening your voice.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    LinkedIn Humanizer: Remove AI Tells

    LinkedIn Humanizer: Remove AI Tells

    Example session with this skill installed

    Here's my draft: "Excited to share that our team has leveraged cutting-edge solutions to revolutionize how startups streamline their workflows. It's not about tools, it's about mindset. We iterate. We learn. We grow. Let me be honest, the journey wasn't easy. Grateful for this amazing team! Thoughts?" Humanize it.

    • Read your context and instructions
    • Compiled the linkedin humanizer

    Scored first. Density markers: leveraged, cutting-edge, revolutionize, streamline (four in the opening lines, one paragraph at 3+ = 2 points). Reveal bridge: "It's not about tools, it's about mindset" (1). Fragment stack: "We iterate. We learn. We grow." — also a parallel triad (2 + 1). Sincerity frame: "Let me be honest" (1). Cliché closer: "Thoughts?" (1). Total 8 → full rewrite pass.

    Rewrite, in order. Leakage: none. Structure: the fragment stack merges into a real sentence; the bridge is deleted and the point stated directly; the closing question is dropped. Vocabulary: the dense opening is rewritten for concreteness instead of synonym-swapped. Proof: here is the honest limit — your draft contains no number, no named thing, and no date, and I will not invent them. Below, brackets mark the two spots where one real specific from you turns this from clean to convincing.

    After: "We spent the spring rebuilding how our startup clients get from signup to first project. [One specific: the step that used to take X days now takes Y.] The tools changed less than the order we do things in — we map the workflow with the team before we automate anything, because automating a broken sequence just breaks it faster. [One dated fact: what happened, when, with one client or cohort.] The team did the unglamorous part well: saying no to the features that sounded good in the demo."

    Recomputed score
    0. What changed and why is the change note you would get with it; what I could not fix — the missing specifics — is flagged rather than fabricated, which is the difference between humanizing a post and writing fiction in your name.

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

    About this skill

    Readers have learned to smell machine writing on LinkedIn, and the platform's low-quality filter quietly throttles posts that match the pattern. The tells are rarely single words; they are patterns at density: three "leverage / robust / landscape"-class words in one paragraph, a reveal bridge drumming up a point that needed no drumroll, a stack of punchy fragments, a second and third perfectly parallel triad, a framed confession announcing honesty instead of being honest, and template leakage no human ever types.

    This skill works a draft in a fixed order. First, three reads — the ear read (would you say it to a colleague?), the fingerprint read (count markers per paragraph), and the intent read (which sentences exist only to sound wise). Markers convert into a severity score with a decision table, so the response is proportional: a clean draft ships untouched, a middling one gets a light pass, a patterned one gets the full rewrite. The rewrite strips leakage, merges fragment stacks, breaks surplus triads, thins density paragraphs by rewriting for concreteness rather than swapping synonyms, and plants real proof — a number with a referent, a named entity, a dated fact — while never inventing specifics the author did not supply.

    The final step is a restraint check against the rewrite itself, because humanizer output has its own fingerprint: manufactured sentence-length seesaws, quirks scrubbed to zero, edits invented to justify the pass. An audit mode produces findings only — blockers, warnings, and the fix for each, in a fixed report shape — for authors who revise their own text. One boundary stated plainly: this fixes what costs you readers, not detector scores, which on short posts are mostly noise.

    This skill edits; it does not draft from scratch, post for you, or guarantee any AI-detector outcome — detector scores on short posts are mostly noise, and any tool promising otherwise is selling the noise. It is calibrated to LinkedIn's norms and English-language posts; other platforms and languages have different tell profiles. And it will not invent specifics: if your draft has no numbers, names, or dates, the rewrite stays honest and flags the gaps for you to fill, which means the last step of a full pass is sometimes yours.

    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 any payment.

    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 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    No promise is made, because none can be kept — on short posts those scores swing wildly on trivial changes. What this fixes is the thing that actually costs you: human readers pattern-matching your post as machine output and stopping, plus the platform's low-quality pattern filter.

    That is the cheapest use case. The three reads plus the score take a minute, and the decision table's first row is "ship it" — the skill is explicitly built to leave clean drafts alone, and says so in the change note.

    Method and restraint. A named tell taxonomy, a scored severity decision, a fixed rewrite order, a never-invent rule for specifics, and a final check that the rewrite did not install the second-generation tells (seesaw rhythm, zero quirks) that generic humanizing usually adds.

    New in Writing & Content