tenfoldmarc/llm-council-skill — reverse-engineered prompt

Build me an installable Claude Code skill called LLM Council. I want to be able to type things like “council this”, “run the council”, “pressure test this”, or “war room this” followed by a messy decision I’m trying to make, and have Claude run it through five different AI advisors instead of just giving me one agreeable answer.

The skill should gather useful context from the current workspace, rewrite my question in a neutral way, then ask five advisors to respond with different mindsets, a contrarian, a first principles thinker, an expansionist, an outsider, and an executor. After that, have them review each other anonymously, then produce a final chairman style verdict that shows where they agree, where they disagree, what blind spots were found, the recommendation, and one practical next step.

Please generate both a nice HTML report and a full markdown transcript, and make the setup simple enough that I can drop it into my Claude skills folder and use it right away.