Growth content orchestrator

Growth Content Orchestrator

The single entry skill for growth and content-marketing work. It locates the task on the funnel stage × asset type map and delegates to one of 13 specialist spokes. The cross-cutting decision every asset shares — searchable vs. shareable, the buyer stage it targets, and the one source-of-truth positioning doc all spokes read first — lives in growth-content-core; read it before planning a campaign or writing an asset cold.

Cluster map (spoke → role)

Foundation (read first)

• product-marketing-context — establishes positioning, ICP, voice, proof. Writes .agents/product-marketing-context.md; every other spoke reads it. Run this before anything else on a new project.

Plan (what to make)

• content-strategy — pillars, topic clusters, keyword-by-buyer-stage, the editorial calendar
• marketing-ideas — the 139-idea bank when the user is stuck or wants channel inspiration
• marketing-psychology — mental models and behavioral principles that inform every other choice

Write (the asset itself)

• content-research-writer — long-form: blog posts, articles, newsletters, thought leadership, with research + citations
• copywriting — conversion copy for pages: homepage, landing, pricing, feature, about
• copy-editing — the seven-sweep polish pass on existing copy (run after a draft)

Convert (turn attention into leads/replies)

• lead-magnets — gated downloadables: checklists, templates, ebooks, swipe files
• free-tool-strategy — engineering-as-marketing: calculators, graders, generators
• cold-email — B2B outbound: prospecting emails + multi-touch follow-up sequences
• email-sequence — lifecycle/nurture flows: welcome, onboarding, re-engagement, drip
• ad-creative — paid ad copy at scale: headlines, descriptions, primary text, iteration from performance data

Launch (ship it)

• launch-strategy — Product Hunt, GTM, the ORB channel model, the five-phase rollout

Picked-up spokes (see dedicated section below)

• avoid-ai-writing — strip AI-isms from a draft (run after copy-editing, before publish)
• zapier-make-patterns — no-code Zapier/Make automation for marketing ops & lifecycle plumbing

Routing rules by intent

The user says…	Route to
"What should I write about?" / pillars / calendar	content-strategy
"I'm stuck, give me growth ideas"	marketing-ideas
"Why do people buy / persuasion / bias"	marketing-psychology
"Write a blog post / article / newsletter"	content-research-writer
"Write/rewrite my landing/homepage/pricing copy"	copywriting
"Edit / tighten / polish this copy"	copy-editing
"Lead magnet / freebie / gated PDF"	lead-magnets
"Build a free calculator/grader/tool for leads"	free-tool-strategy
"Cold outreach / prospecting / nobody replies"	cold-email
"Drip / nurture / welcome / onboarding emails"	email-sequence
"Ad copy / RSA headlines / Meta primary text"	ad-creative
"We're about to launch / Product Hunt / GTM"	launch-strategy
Positioning / ICP / "set up context"	product-marketing-context
"Remove AI-isms / sounds like AI / de-slop this"	avoid-ai-writing
"Automate this in Zapier/Make / no-code workflow / connect these apps"	zapier-make-patterns

Disambiguation:

• New page text → copywriting; improving existing text → copy-editing; emails the page links to → email-sequence.
• Cold (no relationship) → cold-email; warm/lifecycle (already a contact) → email-sequence.
• Downloadable content (ebook, checklist) → lead-magnets; interactive software (calculator) → free-tool-strategy.
• "Marketing ideas" is the starting point when intent is vague; once a channel is chosen, hand to that channel's spoke.

Standard Operating Flow

1. Check the foundation. If .agents/product-marketing-context.md doesn't exist, run product-marketing-context first (or note its absence). Every spoke writes better with it.
2. Locate the task on funnel stage (plan → write → convert → launch) and asset type.
3. Pull the core decision from growth-content-core — every asset must be searchable, shareable, or both, and aimed at one buyer stage. State which before producing.
4. Delegate to the spoke(s). Multi-step asks fan out in pipeline order — e.g. "launch our new feature" → content-strategy (angle) → copywriting (landing page) → email-sequence (onboarding) → launch-strategy (rollout).
5. Return: chosen spoke(s), the searchable/shareable call, the target buyer stage, and the next action.

Guardrails

See growth-content-core. In short: never invent proof — fabricated stats, testimonials, or customer logos erode trust and create legal exposure; pull from product-marketing-context.md or ask. One asset, one job, one CTA, one buyer stage — don't make a page or email do everything. Customer language over company language — mirror verbatim voice-of-customer, not buzzwords ("leverage," "synergy," "best-in-class" are banned). Draft, then copy-editing — don't ship the first pass. Use urgency/scarcity only when genuine.

Picked-up spokes

Standalone skills adopted from the antigravity-awesome-skills library (MIT). They extend the cluster at the edges of the core funnel:

• avoid-ai-writing — audits and rewrites prose to remove 21 categories of AI writing patterns via a 43-entry replacement table, then re-audits the rewrite. Sits after copy-editing as a final de-slop pass before publishing any AI-assisted draft. Route here when the user says "remove AI-isms," "make this sound less like AI," or "de-slop this."
• zapier-make-patterns — no-code automation playbook for Zapier and Make (Integromat): platform choice, trigger-action and multi-step patterns, error handling, and the line at which to graduate to code. The marketing-ops plumbing layer — wiring lead capture, lifecycle emails, and CRM sync without engineering. Route here when the user names Zapier/Make/integromat/zaps/ scenarios or wants to "connect these apps" / "automate this" without writing code.

Loading spokes on demand

To keep CLI startup context lean, this cluster's spokes are not separately registered as skills — only this orchestrator and its *-core are enumerated. When you route to a spoke named above, load it on demand by reading its file:

~/.agents/skill-clusters/skills/<spoke-name>/SKILL.md (or skills/<spoke-name>/SKILL.md inside the skill-clusters repo).