Ezoic Marketing
Ezoic's marketing playbook in code. One entry point routes any marketing, sales, or brand task to a specialist that loads the canonical Ezoic brand system and enforces it end to end — so every output is on-voice, on-positioning, and backed by sourced proof.
Voice: direct, confident, technical. Category: revenue infrastructure for the modern web.
Start here
/ezoic-marketing:start
The orchestrator boots the brand context, runs a short triage, builds a brief, then either drafts trivial things inline or routes to the right specialist. Examples it handles:
- "Draft a LinkedIn post from Tyler on the 250K threshold."
- "Build a one-pager for the SaaS monetization play."
- "Make a deck for an enterprise pitch to a $1M+ web app."
- "We need a landing page for ezID."
- "How do we respond to a publisher who says 'we already have AdSense and it's fine'?"
- "Plan a campaign for the AI Bid Flooring launch."
- "Position our rewarded ads product for online games."
- "Review this draft against the brand voice."
You can also invoke any specialist directly; each re-validates against the brand contract.
Skills
| Skill | Use it for |
|---|
| start (orchestrator) | Entry point. Triage → brief → draft inline or route. |
| ezoic-draft-content | Press release, LinkedIn, X, blog, social, ad copy, FAQ |
| ezoic-collateral | One-pagers, product sheets, case studies, battlecards |
| ezoic-pitch-deck | Sales/pitch deck outline (+ .pptx render when anthropic-skills:pptx is installed) |
| ezoic-landing-page | Landing pages, microsites, product pages |
| ezoic-email-compose | Customer, grandfathered, win-back, nurture, outreach emails |
| ezoic-objection-response | Reframes + proof + rep talk-track for objections |
| ezoic-campaign-strategy | Multi-channel campaign core + rollout matrix + sequencing |
| ezoic-positioning-messaging | Category claim, three-pillar narrative, value-prop ladder, messaging matrix |
| ezoic-brand-review | Section-by-section brand audit with severity-tagged fixes |
How brand truth is resolved
Every skill reads references/_resolver.md first, then resolves brand facts in three tiers:
- MCP — the
ezoic-brand server (read_brand_guide, search_brand_standards, list_brand_assets, get_brand_asset), when approved and reachable.
- WebFetch — the live API on
iheartezoic.com (/llms-full.txt, /api/ezoic-brand, /api/ezoic-brand/assets, …) when MCP is unavailable.
- Local references — the
references/*.md files bundled with the plugin.
If none of the tiers can resolve a fact, the skill stops and asks — it never fabricates. MCP is optional: everything works via the API and local references. See INSTALL.md.
The brand contract (every output passes)
12 always-on rules plus 6 visual rules, defined in references/_resolver.md. The essentials:
- Grep before you write — every product, partner, and stat traces to a canonical source.
- Data-led: every claim backed by a number from
proof-points.md, cited with its segment context.
- Right audience noun — not "publishers" alone when the audience is broader (web apps, games, SaaS, newsletters).
- Category framing is revenue infrastructure / operating layer — never ad network, plugin, or tool.
- No CPM dollar claims without a cited source. Refocus, never attack competitors.
- One narrative frame per asset. Customer is the subject.
- Visual: near-black + off-white carry the system, Mint is the single accent, no all-caps/heavy-bold, no fake dashboards, assets only from the live catalog.
Each skill ends by running the 10-point smell test from brand-voice.md verbatim.
Brand source
Canonical brand library: iheartezoic.com (live page, REST API, MCP, and llms.txt / llms-full.txt). The bundled references/*.md are the brand team's operational layer on top of it.
Install
See INSTALL.md.
