Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Said My Way

    2

    Strip AI buzzwords and match your personal writing cadence using persistent voice profiles and script-based verification.

    Free

    3 usesSecurity scanned
    Said My Way

    Said My Way

    Example session with this skill installed

    Said My Way. Match my voice from this sample: "We need to ship the API docs by Friday or the integration team will be blocked. No excuses." Draft a reply to the team lead.

    • Read your context and instructions
    • Compiled the said my way

    I have analyzed your sample and updated your voice profile. The draft below follows your direct, urgency-driven cadence and avoids the standard AI greeting sandwich.

    We have to get the API documentation out the door by Friday. If we miss this window, the integration team hits a wall and the whole schedule slips. I am prioritizing the endpoint descriptions today to stay on track. Let's clear the blockers now so we can finish this.

    TELL
    Said: "We need to ship the API docs by Friday or the integration team will be blocked."
    Cut: The integration of these components is a multifaceted task that requires our holistic attention.
    Put: If we miss this window, the integration team hits a wall and the whole schedule slips.

    Next steps

    • Review the voice-card.md file created in your directory to see the captured habits.
    • Ask for a "memo" or "email" if you need to bypass the 60-word social media length limit.
    • Run "wipe card" at any time to reset to the default plain prose settings.

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

    About this skill

    AI-generated content often sounds like a generic press release. It uses the same "tapestry" of "robust" buzzwords and "delves" into topics with a predictable, mechanical cadence that screams machine-written. This skill stops that by forcing the LLM to adopt your specific writing patterns, sentence structures, and vocabulary constraints while actively stripping out AI cliches.

    What it does

    • Eliminates AI-isms by banning a specific list of 20+ common LLM filler words like "delve," "myriad," and "seamless."
    • Matches cadence by learning your specific sentence habits and varying lengths to avoid repetitive clause structures.
    • Enforces word counts using a Python script to ensure social replies stay within a tight 45 to 60-word window.
    • Learns your voice by generating and maintaining a voice-card.md file based on samples of your actual writing.
    • Identifies the tell by quoting a line you actually said and replacing the draft sentence that feels the least like you.

    How it works

    1. Intake triggers when you provide writing samples or ask it to "match me."
    2. Analysis creates a voice-card.md file in your directory to store your specific sentence habits and banned phrases.
    3. Drafting applies these constraints to your request, filtering for plain American prose and varying sentence structures.
    4. Refinement runs a count script and a "Tell" check to swap out the most AI-sounding sentence for a better match.

    Frameworks & tools

    Uses a local Python environment to run scripts/count.py. Stores user profiles in a local voice-card.md file within the working directory.

    Why this beats prompting it yourself

    Most prompts to "sound human" just result in more cliches or overly casual slang. This skill uses a persistent data file and automated script checks to enforce structural rules that simple prompting can't maintain over long sessions.

    Use cases

    • Drafting LinkedIn or X posts that don't look like they came from a prompt.
    • Rewriting generic AI summaries into your specific professional tone.
    • Cleaning up internal memos to remove "corporate-speak" and filler.
    • Standardizing writing styles across a small team using a shared voice card.

    Known limitations

    Requires a Python 3 environment to run the word count script. Will not perform tasks like coding, recipes, or general Q&A; it focuses strictly on stylistic drafting.

    Requires Python 3 for the count script. Card storage is local to the working directory. Focused on drafting; skips general Q&A or code tasks.

    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

    3 uses

    Times used to date

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 26 days ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed1 month ago
    Updated26 days ago

    What's inside

    Frequently Asked Questions

    This skill solves the "AI fingerprint" problem by actively banning specific cliches (like 'delve' or 'tapestry') and using a script to enforce your specific sentence cadence, ensuring content doesn't sound mechanical.

    You will receive the core AI skill instructions, the `scripts/count.py` file for precise length enforcement, and a template for the `voice-card.md` profile system.

    Yes, this skill requires a local Python 3 environment to run the word count and cadence verification scripts, as well as the ability for your AI agent to read and write to a local `voice-card.md` file.

    Simply provide a few samples of your writing and use the "match me" command; the skill will automatically analyze your style and generate a persistent `voice-card.md` file to guide all future outputs.

    While the skill is optimized for plain American prose, you can manually edit your `voice-card.md` file to include or exclude specific regional vocabulary and industry-specific terminology.

    This skill is designed specifically for stylistic drafting and rewriting; it will not perform general tasks like writing code, providing recipes, or handling complex logic outside of content creation.

    New in Productivity

    Free