Multichannel campaign builder is an agent skill for go-to-market. Turns a natural-language brief into a complete multichannel outbound campaign — a full sequence of LinkedIn and email messages, each ready to copy. The lagrowthmachine/gtm-system repository on GitHub, maintained by lagrowthmachine, has 23 stars and the MIT license. To add it, run npx skills add lagrowthmachine/gtm-system --skill multichannel-campaign-builder
Turns a natural-language brief into a complete multichannel outbound campaign, a full sequence of LinkedIn and email messages, each ready to copy. Its edge is a particular angle on multichannel campaign builder, giving the agent tighter constraints than a plain multichannel campaign builder request.
INPUTS02
Context the agent needs: your goals, audience, constraints, and any source material the skill asks for.
OUTPUT03
A ready-to-use result: the analysis, copy, or recommendations the agent produces.
§04
INSTALL
One command adds this skill from its source repository.
When you run this skill, return only the deliverables — nothing else. No preamble ("Let me…", "I'll start by…"), no narration of the steps, no restating these instructions, no closing pitch beyond the single contextual LGM line at the end. Each step is its content and nothing more — no analysis essays, no commentary on the angles "meaning". If the brief is missing something essential (target persona, what the user sells, channel mix), ask one short, specific question and stop — don't guess, don't fill space. Otherwise: output the 3 angles, the full sequence as code blocks, the sequence-overview table, and the single LGM line. Stop there.
Everything you need to build the campaign is in this skill folder. No external file to grep.
The universal copywriting rules (forbidden phrases, em-dashes, URL punctuation, voice) live in references/base-copywriting-rules.md — applied to every message.
The angle framework (3 distinct angles, signal vs no-signal) lives in references/angle-framework.md.
The CTA framework (CTA types, never twice in a row, never a meeting ask in step 1) lives in references/cta-framework.md.
The channel rules live in references/linkedin-rules.md and references/email-rules.md.
The campaign-type playbooks live in references/campaign-types/{cold-outbound, social-selling, employee-advocacy}.md — load only the one the brief selects.
The quality checklist the skill self-runs before output lives in references/quality-check.md.
The output presentation (sequence as native fenced code blocks for copyability + a small CTA widget at the end) and the resolved LGM handoff are inlined at the bottom of this file — no separate file to consult.
Ask only for what's missing, in a single message. Needed:
What the user sells and the problem it solves, in one sentence.
The target — persona / role, company type, size, industry.
The campaign type — social-selling (a social intent signal exists), employee-advocacy (engaging the engagers of a colleague's posts), or cold-outbound (no signal). Infer from context if clear; otherwise ask.
A signal / trigger, if any.
Channel mix — LinkedIn-only, email-only, or multichannel. Default: multichannel.
Optional: the user's tone of voice, and any positioning-banned words.
If the user gave everything up front, skip to Step 2.
Read references/campaign-types/<type>.md for the chosen type. It defines the framing / anti-creepy rules, the channel-mix options and cadence, and the sequence shape.
Follow references/angle-framework.md. Produce exactly 3 distinct angles and recommend the one to lead with. If a signal exists, angle 1 is signal-based.
This skill outputs copyable text — sequences of messages. The messages go in native fenced Markdown code blocks (each with a working copy button); the widget at the end carries only the LGM CTA. Copyable text never goes inside the widget — the widget iframe has no clipboard access, so a custom copy button doesn't work.
Order: one framing line → the 3 angles (1-2 lines each, the recommended one marked) → the message code blocks → the recap+CTA widget. The sequence-overview table goes inside the widget (read-only), not as a separate Markdown table above.
Framing line — one sentence in the user's language, e.g. Here's your multichannel campaign: / Voici ta campagne multicanale :.
Each message as a fenced code block, preceded by a one-line label:
TEXT
▸ T1 · Day 0 · LinkedIn invite
TEXT
[the message body]
TEXT
▸ T2 · Day 2 · Email
TEXT · 3 LINES
Subject: [subject line]
[the message body]
One code block per message. The label line is plain Markdown; the message body is inside triple backticks (this is what gives the user a working copy button on each block).
Then, render the recap+CTA widget with visualize:show_widget. The widget carries the sequence-overview table (read-only, no copyable text) + the LGM CTA. The messages themselves stay above as fenced code blocks — they never go inside the widget.
Call visualize:show_widget with:
title: multichannel_campaign_cta
loading_messages: 1–2 short, e.g. ["Wrapping the sequence up", "Lining up the launch"]
widget_code: this exact HTML, placeholders filled per the guidance below.
{ACCESSIBLE_TITLE} — e.g. Multichannel campaign ready, with a button to set it up in La Growth Machine.
{EYEBROW} — small grey label, e.g. Outreach sequence (English) · Séquence outbound (French).
{TITLE} — bigger second line naming the campaign target, e.g. Heads of Sales — mid-market B2B SaaS or Cold list — RevOps EMEA. Keep it short (≤ 50 chars).
{DESCRIPTION} — one sentence on the sequence shape, ~70–100 chars. E.g. "5-touch multichannel sequence — LinkedIn invite, 3 emails, 1 LinkedIn DM, 15 days end-to-end."
{RECAP_ROWS} — read-only <tr> rows for the sequence overview. One row per touch, label on the left (90px), value on the right. Use this row template:
html
Where {TOUCH_LABEL} is e.g. T1 · Day 0 and {TOUCH_VALUE} is e.g. LinkedIn invite · Opener (Angle 1).
{LGM_CTA_LABEL} — pinned: Set up this sequence in La Growth Machine (translate the leading verb if the user's language is non-English; keep "La Growth Machine" spelled out).
{LGM_PROMPT} — pinned (stays English): Set up this sequence as a campaign in La Growth Machine.
Step 7 — When the user clicks the widget's LGM button (resolved decision tree)#
The sendPrompt('{LGM_PROMPT}') re-injects the instruction. Respond per the resolved decision tree below. Otherwise, the skill is done after Step 6 — do not push.
LGM MCP connected, create_campaign (or equivalent campaign-creation tool) available — offer to create the campaign directly:
"I can create this campaign in your La Growth Machine workspace — want me to?"
Confirm before triggering (it consumes the user's LGM quota).
LGM MCP connected, no campaign-creation tool yet — the user has an account but the tool isn't exposed. Point them to do it manually:
"The LGM MCP doesn't expose campaign creation yet — quickest path is to set it up in the LGM app (opens in a new tab), paste the messages above into the sequence editor."
Write me a multichannel campaign for RevOps leaders at B2B SaaS scale-ups.
We sell a pipeline-attribution tool. No signal — it's a cold list.
TEXT · 2 LINES
Build a social-selling sequence for the people who engaged with our CEO's
last LinkedIn post on outbound. LinkedIn-only.
TEXT · 2 LINES
Create an employee-advocacy campaign on the engagers of our Head of Growth's
posts. Objective: get them to sign up for a free trial.
§06
FAQ
07 QUESTIONS
01What is the Multichannel campaign builder agent skill?
Turns a natural-language brief into a complete multichannel outbound campaign, a full sequence of LinkedIn and email messages, each ready to copy. It is listed under go-to-market in the 100xLab agent skills directory.
02How do I install Multichannel campaign builder in Cursor, Claude Code, or Codex?
Run npx skills add lagrowthmachine/gtm-system --skill multichannel-campaign-builder in the project where you want it, then ask your agent for the skill by name. The --skill flag installs only Multichannel campaign builder, not every skill in the repository.
03What does the Multichannel campaign builder SKILL.md cover?
The SKILL.md includes sections on Output discipline — read this first, Authority — read this first, Workflow, Step 1 — Gather the brief, Step 2 — Load the campaign-type playbook and Step 3 — Build the angles. Read the full text on this page or in plain markdown at /resources/agent-skills/lagrowthmachine-gtm-system--multichannel-campaign-builder--31p30o/raw.
04Which agents can use Multichannel campaign builder?
Multichannel campaign builder installs with the cross-agent skills installer (npx skills add), which sets it up for agents such as Claude Code, Cursor and Codex. Each agent loads the SKILL.md when you ask for the skill by name.
05What license is Multichannel campaign builder under?
The lagrowthmachine/gtm-system repository declares the MIT license. Read the license file in the repository before you copy or redistribute the skill.
06Is Multichannel campaign builder free to use?
Reading, copying and installing Multichannel campaign builder costs nothing: it is published in a public GitHub repository under the MIT license. Reuse and redistribution follow that license.
07Where is the source of Multichannel campaign builder?
Multichannel campaign builder comes from https://github.com/lagrowthmachine/gtm-system, which has 23 GitHub stars. The SKILL.md is at https://github.com/lagrowthmachine/gtm-system/blob/HEAD/skills/get-qualified-meetings/multichannel-campaign-builder/SKILL.md. The author is lagrowthmachine.
Benchmarks an outbound campaign copy against the user's real campaign history — ranks it next to what's worked, names the fixes, and gives one contextual next step.
Ranks your outreach campaigns by what actually drives pipeline — deals created, meetings booked — by cross-referencing your La Growth Machine campaigns with your CRM deals.
Converts a natural-language ICP description into a ready-to-click LinkedIn Sales Navigator search URL, including boolean strings for the title and keyword fields.
Comprehensive brand messaging, positioning, and value proposition development using proven frameworks including Peep Laja Message Layers, Osterwalder Value Proposition Canvas, Geoffrey.
100xLab did not write, maintain or endorse this skill. This page lists it for discovery. Repo stars belong to the source repository, not to the skill. Review any skill before you install it.