My submission to Peec AI MCP Challenge (I think it is almost too powerful to publish...)

For the Peec AI MCP Challenge, I built 9 Claude skills that sit directly on top of the Peec AI MCP.

One entry point (/peec-start), persistent state, all open on GitHub.

 Three things that genuinely surprised me during the live run:

 → peec-checkup analyzed 994 chats and scored my setup at 56%. Clear weakness: the awareness stage. Data I'd usually have to stitch together by hand.

 → peec-content-intel combines Peec visibility with Visibly AI (GSC, backlinks, query fan-out) and scrapes Reddit + Gutefrage. Output: concrete Reddit threads + AMA recommendations — not a generic list.

 → peec-learn adds a memory layer (SkillMind, Pinecone-ready) across the entire project. Lessons from project A inform project B, cross-client.

My take: tools like Peec AI only unlock their full value once you pull them out of the dashboard and embed them into an agent loop.

Full video + repo: https://lnkd.in/dwQ4_tVy

Lets let the judges descide!
Lily Ray, Malte Landwehr, Ethan Smith,

PS: Support my github with stars:
https://lnkd.in/dAMnxBQE
#PeecAI #AISEO #ClaudeSkills #MCP #SEOAutomation #builtwithpeec