Theo7777/gtm-skills

13 skill entries

• ad-description-maker

  Write ad descriptions that convert using the Problem to Solution to Benefit formula, in 35 words or fewer, with platform character limits for Google Ads, Meta, and LinkedIn. Use when the user asks for ad descriptions, ad body copy, primary text, introductory text, RSA descriptions, product descriptions, or the copy that sits under a headline. Trigger phrases include "write an ad description", "body copy for this ad", "Meta primary text", "Google ad descriptions", "LinkedIn intro text", and "product blurb". Pairs with ad-headline-maker for the headline above it.

  Self-containedInstructions only1,348 words
• ad-headline-maker

  Write ad headlines that convert using the Core Benefit + Objection Crusher formula, in 10 words or fewer, with platform character limits for Google Ads, Meta, and LinkedIn. Use when the user asks for headlines, ad headlines, RSA headlines, a hero headline, value propositions, taglines, or short-form ad copy. Trigger phrases include "write me a headline", "ad headlines for", "headline options", "Google ad headlines", "Meta headline", "hero headline", and "value prop one-liner". For the body copy under the headline, use ad-description-maker.

  Self-containedInstructions only1,176 words
• ad-message-maker

  Turn a business into three static ad variations built as editable frames in Figma. Same hero photo and layout, three different messages (pain, desire, proof), so you test the message and not the design. Reads the shared GTM context or whatever the project folder already has, and asks five questions only when there is none. Works in Claude Code and Codex with the figma-console MCP; without Figma it still writes the copy and image prompt. Use when the user wants static ads for a product or service, or wants to test messages. Trigger phrases include "ad in Figma", "3 ad variations", "hero ad", "figma ad", "ad creative in figma", "test three messages", and "ad message maker".

  Needs a keyDeclared tools2,346 words
• cold-outreach

  Write one signal-driven cold email, LinkedIn connect note, or LinkedIn message for a named person. Takes what you sell, who you're writing to, and one reason to reach out now (from a library of 41 signals, or your own trigger), and returns a short message built on Hook, Why, Objection, CTA. Use when the user says "cold email for", "write outreach to", "LinkedIn connect note for", "LinkedIn message for", "signal email", or names a person, a company, and a reason to reach out. For multi-step sequences, use an email sequence skill instead.

  Reaches the webInstructions only952 words
• deep-ad-research

  Research what video ads are working in any niche. Scrapes the Meta Ad Library for competitor or keyword ads through Apify, transcribes them with OpenAI Whisper, sorts longest-running first, then writes an insights report on hook patterns, angles, offers, and what it means for your brand. Use when the user wants to see what competitors run on Facebook or Instagram, or which ad hooks and angles dominate a market. Trigger phrases include "research competitor ads", "what ads are working in my niche", "Meta ad library research", "spy on competitor ads", "what is [competitor] running", and "deep ad research".

  Needs a keyDeclared tools1,423 words
• gtm-context

  Builds and maintains the shared GTM context file (the "GTM brain") at .agents/product-marketing.md that every other skill in this pack reads, so the user never has to explain their product, customer, and positioning twice. Use at the start of any new project, or when the user asks to "set up context", "set up the GTM brain", "create a product marketing context", "describe my product", "who is my ICP", "update our positioning file", "add this to our context", or shares a website, docs, or repo and wants the foundations captured. Also use to update one section, reconcile conflicting information, or migrate an older context file. Reads and extends an existing product-marketing file rather than overwriting it.

  Reaches the webInstructions only1,524 words
• icp-scoring

  Score a list of target accounts against an ICP using a fit gate and weighted buying signals, then band them A Hot / B Warm / C Nurture / D Monitor with a recommended action and the reasons behind every score. Accepts a pasted list, a CSV, or a spreadsheet export. Use when the user asks to score accounts, prioritise a target list, rank prospects, qualify leads against the ICP, tier accounts, or apply ICP or signal scoring.

  Self-containedInstructions only878 words
• jtbd-research

  Jobs-to-be-done research on any problem space. Searches Reddit, review sites, forums, and community threads (and any interview transcripts you supply) for real customer language, then extracts job statements, ranked pains, loves and complaints about current tools, switching triggers mapped to the forces of progress, and a verbatim quote bank with sources. Use when the user asks for JTBD research, jobs to be done, voice-of-customer research, audience pain research, switching triggers, why people switch, or real quotes to use in copy.

  Self-containedInstructions only977 words
• linkedin-connection-analysis

  Turn a LinkedIn connections export (Connections.csv, with name, company, position, and date connected) into a simple prioritised network CRM for a stated goal, such as finding buyers, referrers, hires, investors, or advisers. Scores each person High/Medium/Low fit against the goal and the ICP, keeps only Medium and High, and sorts best fit to the top. Keeps personal data local by default. Use when the user shares a LinkedIn connections export, Connections.csv, or asks "who in my network can help with", "mine my LinkedIn connections", "warm intros", or "who should I reach out to first".

  Reaches the webInstructions only847 words
• linkedin-post-distill

  Distil a LinkedIn post into a one to three-word phrase naming its core topic, using the author's own words. Use when the user pastes a LinkedIn post and asks what it is about, wants a topic label, a tag, or a one-line summary, or says "distil this", "what's the core message", "label this post", or "what is this post on". Also use in an outreach or engagement loop to tag posts that prospects have written or engaged with, so an opener can reference them. Output is the phrase only, formatted to fit "{author}'s post on {phrase}".

  Self-containedInstructions only733 words
• normalise-outreach-names

  Normalise first names and company names in an outreach list so template merge fields read naturally. Strips emoji, credentials, honorifics, and legal suffixes, shortens companies to the name a person would say aloud, and enforces a caps policy (no ALL CAPS unless it is a true acronym or the brand styles itself that way everywhere). Use before any CSV goes into an email tool or enrichment table, or when the user says "normalise names", "clean up the company names", "fix the first names", or "the merge fields look wrong".

  Self-containedInstructions only769 words
• positioning-messaging

  Positioning and messaging for B2B products and websites, worked top-down from positioning to messaging to copy. Use when the user asks for website copy, homepage messaging, a hero section, benefit sections, a value proposition, a positioning statement, messaging building blocks, or landing page copy, or says things like "what do we even say we are", "who is this for", "why are we better", "rewrite our homepage", or "audit this page". Also use when reviewing or auditing existing website copy. Covers strategic foundations, full homepage copy (hero to footer), and individual sections.

  Self-containedInstructions only2,029 words
• tech-stack-prospects

  Build a prospect list from a technology signal. Finds companies that run a named tool (TheirStack technographics), filters them by how long they have had it, how recently it was seen, and how confident the detection is, then finds the right people at each company with a work email (Apollo) and writes the list to a CSV. Can also load the list into lemlist as a draft campaign. Use when the user says "find companies using X", "who runs X", "prospects on a competitor's tool", "technographic list", "find people at companies using", "switch campaign", or names a tool plus the kind of company they sell to.

  Reaches the webActs undeclared1,993 words