esmefong/gtm-skills

7 skills

• account-sourcing

  Turn an ICP into a ranked target-account list with receipts. Takes the account-level ICP (ideally produced by icp-research), translates its criteria into queries against public sources (job postings, press, funding databases, registries, tech lookups), builds the account universe, scores and visibly disqualifies against the ICP rubric, and tags each surviving account with a dated why-now signal. Every account carries public evidence; no account is invented. Use whenever the user says "build a target account list", "account sourcing", "TAM list", "list building", "find companies to sell to", "who should we go after", "turn our ICP into a list", "prospect list", "account universe", or "lead sourcing". Ships with three key-free scripts: a deterministic account scorer, a live hiring-signal detector for public ATS job boards, and an executive briefing renderer. For products bought by organizations; consumes icp-research output and feeds outbound-engine.

  Reaches the webActs undeclared1,225 words
• brand-voice-guide

  Extract a brand's voice from its real writing and build a structured voice directory a team (or an AI agent) can write from consistently. Harvests existing content — homepage, blog, posts, emails, changelog — into voice-core rules, per-channel guidance, and a writing standard whose Avoid section names the banned words and voice-killing patterns, each grounded in verbatim samples. Interviews the user only to fill what the samples can't answer. Use whenever the user says "brand voice", "voice guide", "tone of voice", "voice and tone", "brand voice guide", "define our voice", "extract our voice", "style guide", "voice standard", or wants to document how their brand sounds so content stays on-brand at scale. Works across industries and channels. Builds the guide; it does not police drafts — that is the job of a separate content-quality-check skill.

  Reaches the webActs undeclared1,616 words
• content-quality-check

  Editorial QA pass that catches AI tells, template language, hollow hedges, weak closers, and brand-voice violations in any draft before it publishes. Runs a deterministic linter (slop phrases, LLM-overrepresented words, "not just X, it's Y" reframes, em-dash chaining, rhetorical-question openers, announcement cliches, the brand's own Avoid list) and then an editorial judgment pass (throat-clearing openers, closers that restate the open, claims without proof, specificity, voice-pillar conformance). Reports findings by severity with line references and a fix per flag; rewrites only on request. Use whenever the user says "review this draft", "QA this post", "check for AI tells", "does this sound AI-written", "editorial review", "content quality check", "de-slop this", or is about to publish copy. Enforces the standard that brand-voice-guide builds; works standalone with the universal checklist when no voice directory exists.

  Self-containedActs undeclared749 words
• icp-research

  Turn raw research signals into an operational Ideal Customer Profile, in any industry. Ingests interview transcripts, survey exports, community/Slack threads, review sites, sales-call notes, and funnel data, then produces an ICP document built on situation triggers and a how-to-win plan. Use whenever the user says "ICP research", "ideal customer profile", "build a buyer persona", "profile our target buyer", "who should we sell to", "audience research", "customer segmentation", "voice of customer", or asks to turn interviews / community threads / reviews into personas. Works for consumer/lifestyle, B2B SaaS, fintech, and enterprise alike: the situation-first core is industry-neutral, and it adds a buying-committee map and account-level scoring rubric only when the buyer is a team. Built on the 5 Rings of Buyer Insight (Revella), Jobs-to-be-Done and the Four Forces (Moesta / Christensen), and demand-side selling.

  Reaches the webInstructions only2,265 words
• market-prioritization

  Rank candidate markets, segments, verticals, or geographies against weighted criteria and report how robust the ranking is. Builds a criteria model from strategy (opportunity, fit, access, economics, risk), scores each market with evidence per cell, runs sensitivity analysis to show which rankings survive a reasonable argument about the weights, and names the deprioritized markets with reasons. Use whenever the user says "market prioritization", "which market should we enter", "segment prioritization", "vertical strategy", "expansion strategy", "TAM by segment", "where should we focus", "scoring model", or has more opportunities than resources. Consumes icp-research where available and feeds account-sourcing, which sources accounts inside the markets this skill picks.

  Reaches the webActs undeclared1,142 words
• outbound-engine

  Turn a sourced account and its live signal into outreach worth sending, or refuse with reasons. Consumes account-sourcing output (a qualified account with a dated signal, an owner, and an entry role), the ICP from icp-research (triggers, buying committee, how-to-speak rules), and the voice directory from brand-voice-guide, then drafts a committee-aware sequence in the brand's voice with per-line provenance. A mechanical no-send gate rejects weak sends: stale signals, weak signal-to-trigger mapping, banned voice, fake familiarity, missing asks. Use whenever the user says "outbound", "cold outreach", "write to this account", "draft a sequence", "outreach sequence", "ABM sequence", "signal-based outreach", "cold email", or wants sourced accounts turned into messages. Role-level targeting only: it never harvests individuals or contact details.

  Reaches the webActs undeclared1,138 words
• positioning-message-house

  Build a positioning message house for a company: one falsifiable category bet, two to three message pillars, and a translation matrix that carries the same messages across go-to-market functions (marketing, sales, product, customer success) at their buyer-journey moments, plus a generation system that emits battlecards and objection-handling docs from the structured house. Consumes the ICP from icp-research (triggers, committee, decision criteria, how-to-win) and the voice from brand-voice-guide. Use whenever the user says "message house", "messaging framework", "positioning framework", "messaging matrix", "core narrative", "category narrative", "align our messaging", "battlecard", or wants one story translated across teams instead of four teams inventing four stories.

  Reaches the webActs undeclared833 words