Works with the AI tools you already use

    CClaude CodeCCursorCCodex CLIGGitHub CopilotGGemini CLIVVS CodeWWindsurfMManus+14 more

    DeskRoot Marketing Pack: five skills, five worked examples

    by DeskRoot

    1

    Requires DeskRoot and a paid Claude plan with Cowork, both sold separately. Five marketing skills, each with a full worked example you can read before you rely on it.

    Secure checkout via Stripe

    0 installsSecurity scanned

    See it in action

    You say

    Develop a positioning statement for our new database tool comparing it to MongoDB, and list which features we should omit from the messaging.

    Your agent does

    Positioning Statement: [Substantiated] Faster query execution for time-series data vs MongoDB. Omitted Features:

    • Document store flexibility (Reason: MongoDB parity, not a differentiator).
    • Mobile sync (Reason: Unsubstantiated performance claims).

    What you get

    Enforce one primary metric and a kill criterion for every marketing campaign.Synchronize product claims across emails, landing pages, and changelogs.Analyze search competitors to find gaps before writing new content.Review campaign performance without inflating ROI via unattributed conversions.

    About this skill

    This pack requires DeskRoot, which is a separate purchase, and DeskRoot requires a paid Claude plan with Cowork. Neither is included here and both cost money. On its own this is twelve documents with nowhere to live. DeskRoot is the workspace; these are five skills that install into its skills library.

    WHAT IS IN IT

    Positioning statement. Anchored to a named alternative your buyer is actually choosing between, because positioning that compares your product to nothing is true of everything. Every claim in the output is marked substantiated or not, and the features that were cut are listed with the reason they were cut.

    Campaign brief. One primary metric, a stated budget, a decision date, and a kill criterion written as a number and a date before the plan is written rather than after it fails.

    Launch announcement set. Every surface written in one pass from a single claim ledger, so the email, the site page, the posts and the changelog cannot disagree about the price, the date, or what actually shipped.

    Search content brief. Built from the pages already ranking rather than from a keyword list, with a section naming what your page will contain that the current results do not.

    Campaign performance review. Grades the campaign against the target it was actually given, quoted from the brief.

    WHAT MAKES THESE DIFFERENT IS WHAT THEY REFUSE TO DO

    A language model asked to write positioning will produce fluent copy from nothing at all, and fluent copy built on no comparison is the most common failure in marketing work. These skills are built to stop rather than to fill the gap.

    The positioning skill will not draft until you name the alternative. The campaign brief will not accept awareness as a primary metric unless you supply the instrument that measures it. The launch skill strikes any claim you cannot source and leaves the strike visible rather than deleting it quietly. The search brief will tell you not to write the page if your honest answer to what it will contain that the ranking pages do not is nothing. The performance review gives unattributed conversions their own named line and never spreads them across your channels, because spreading them makes every channel look profitable.

    EVERY SKILL SHIPS WITH A WORKED EXAMPLE

    Each example is a real input paired with the real output the skill produces, on fabricated demo data. Nothing is trimmed or touched up for the page. You can read the output quality before you rely on it, which is why they are in the download rather than in the marketing.

    WHAT YOU ACTUALLY GET

    Twelve markdown files, 79,657 bytes. Five skills, five worked examples, a root instruction file and a short start guide. Nothing in it can run, make a network request, or read anything outside the folder you connect.

    HOW IT INSTALLS

    Drop the five skill files into your DeskRoot skills library. Your agent reads them the same way it reads everything else in the workspace, which means it applies them with your standing rules and your context already loaded. That is the difference between a skill that already knows your product and your buyer, and one that asks you every time.

    DO NOT BUY THIS IF

    You do not have DeskRoot, or you do not have a paid Claude plan with Cowork. Both are required, both are separate purchases, and without either one this will not work. A refund costs us both more than the sale is worth.

    You want software that connects to your ad platforms or your analytics. These are instructions your agent follows, not an integration. Where a skill needs performance or competitive data, you supply it.

    You want the writing done without the thinking. Four of these five skills will stop and ask you for something before they produce anything.

    PERMISSIONS

    None. This download is twelve markdown files. It requests no permissions, has no executable content, makes no network calls, and reads nothing outside the folder you place it in.

    How to install

    Drop the file into your AI Agent. Works with Claude, Cursor, ChatGPT, and 20+ more.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    30-day refund

    Not a fit? Get your money back

    Trust & safety

    Security scanned

    Verified clean today

    Listedtoday

    Creator

    DeskRoot
    DeskRoot

    3 skills on Agensi

    DeskRoot is a structured operating system for Claude. Install it once into Claude and it starts every session already knowing your work. Everything ships as plain markdown you can read before you connect anything.

    Frequently Asked Questions

    Popular in Marketing & SEO