Marketing Skills v2.0.1

A content patch — no new skills, no breaking changes. Three skills refreshed to current best practices, plus four community vendor PRs rolled in.

Big thanks to the community contributors in this release: @berset, @abhibavishi, @tgonzalezc5, and @SalvatoreApicella0 🙌

Updated skills

ai-seo — aligned with Google's official AI optimization guide

Following Google's AI features optimization guide, the skill now:

• Frames Google's official stance vs. how other AI engines (ChatGPT, Claude, Perplexity) behave differently — they reward extractable structure even when Google says it's not required
• Adds Query Fan-Out — Google's documented behavior of generating concurrent related queries under the hood; reframes content strategy from per-query targeting to topical cluster coverage
• Adds Agentic Experiences — browser agents accessing sites via visual rendering, DOM inspection, and accessibility tree; semantic HTML matters more than ever; mentions the emerging Universal Commerce Protocol (UCP)
• Adds "What NOT to Do" — Google's explicit guidance: don't chunk content for AI, don't write separate AI content (risks scaled-content-abuse policy), don't pursue inauthentic mentions, don't block AI crawlers if you want citation
• Sets Search Console expectations — no AI-specific reports exist; standard SEO metrics + third-party tools are the only measurement path
• Adds local/ecom guidance — Merchant Center feeds, Google Business Profile, and Business Agent for AI Search in those verticals
• Reframes (kept) llms.txt + /pricing.md as "Google says not required, but helps non-Google AI engines"

image — refreshed model lineup for May 2026

Latest models now surfaced:

• Gemini Image family with Nano Banana / Nano Banana Pro naming
• Flux Pro 1.1, Kontext (in-image editing), Dev, Schnell variants
• Ideogram 3.0
• ChatGPT Images 2.0 / GPT Image (DALL-E 3 deprecated)
• Midjourney v7
• Recraft V3 (vector + brand illustration)
• Stable Diffusion 3.5 / SDXL

Updated decision tree and trigger phrases.

video — refreshed model lineup for May 2026

Latest models now surfaced:

• Veo 3 (Google) — top quality + synced audio
• Sora 2 (OpenAI) — promoted from limited-availability caveat
• Runway Gen-4
• Kling 2.5 / 3.0 (Kuaishou)
• Seedance (ByteDance) — fast, low-cost batch
• Hailuo / MiniMax — character consistency across shots
• Pika 2.x
• Hunyuan Video / Wan 2 — open-weight self-hosted

New "Quick picks" guide by use case (cinematic + audio / batch / character consistency / self-hosted / storyboard / image-to-video).

New tool integrations (community PRs)

• Cogny (#275) — hosted MCP gateway for marketing channels (Search Console, LinkedIn Ads, TikTok Ads, Plausible, etc.) behind one URL with managed OAuth. Thanks @berset!
• RankParse (#304) — agent-friendly SEO tool: backlinks, domain authority, tech stack at lower per-call cost than enterprise tools; CLI + integration doc. Thanks @abhibavishi!
• Exa (#266) — AI-powered neural web search for LLMs and agents; useful for content research, competitor discovery, news monitoring; CLI + integration doc. Thanks @tgonzalezc5!
• ChatGPT Images 2.0 added to ad-creative static-image tool list (#269). Thanks @SalvatoreApicella0!

Upgrade

No reinstall required. The three updated skills (ai-seo, image, video) will report 2.0.1 in their metadata.version; the other 37 skills remain at 2.0.0. marketplace.json is bumped to 2.0.1.

If your agent checks VERSIONS.md for updates, it'll see these three rows changed and prompt you accordingly.

Full changelog

• #311 — image + video model refresh
• #313 — ai-seo Google guide alignment
• #314 — v2.0.1 version bumps
• #275 — Cogny integration (@berset)
• #304 — RankParse integration (@abhibavishi)
• #266 — Exa integration (@tgonzalezc5)
• #269 — ad-creative tool names update (@SalvatoreApicella0)

Full diff: v2.0.0...v2.0.1

Contributing

If you'd like to add a tool integration or improve a skill, see CONTRIBUTING.md. Tool integrations (a tools/integrations/<name>.md + registry row) are the fastest path to a merged PR.