cohort-claude-support

The money-and-account lane of Cohort Claude Support. Answers every plan, pricing, usage-limit, model, seat, billing, and data-privacy question — and NEVER from memory: every number, plan feature, and policy specific is fetched live from the official pages first, then explained in plain words with a cohort-fit recommendation. Also the honest "what does this whole system cost me monthly" answer including the optional bring-your-own tools. Routes actual billing/login problems to official support (never touches credentials or payments). Trigger on: "which plan do I need", "Pro or Max", "what does Claude cost", "I hit my usage limit", "what are the limits", "which model should I use", "what's fast mode", "can my assistant use my account", "seats", "is my data private", "does Anthropic train on my data", "billing", "I got charged", "cancel / upgrade / downgrade", or any money, plan, model, or privacy question. Usually reached through support-navigator.

The program lane of Cohort Claude Support — everything about the cohort itself rather than the tools: what week the member is in and what that week is for, where the recordings, community, and office hours live, what the whole stack honestly costs, and the calm catch-up plan when they've fallen behind. Reads the cohort knowledge base and the member's start date; where a program fact isn't wired up yet it says so honestly and offers the interim door — it NEVER invents links, dates, or curriculum. Trigger on: "what week am I in", "what's this week", "where are the recordings", "office hours", "community link", "I'm behind", "help me catch up", "what do I need to buy", "what does this all cost", "who do I ask about the program", or any cohort-program question. Usually reached through support-navigator.

The human-handoff lane of Cohort Claude Support. When support can't fix it — two failed attempts, an Anthropic account/billing issue, a real bug, a blocked work account, or a feature request — this skill packages EVERYTHING the session already learned into a complete, finished ticket and routes it to the right human door: Mike's support portal (the ticket text handed ready to paste — the member opens the portal, pastes, submits), Anthropic's official support for account/billing, the cohort community for anyone-else-hit-this questions, or office hours for coaching. The member never writes a support ticket from scratch and never explains their problem twice. Trigger on: "talk to a human", "contact support", "file a ticket", "report a bug", "this still isn't working", "escalate this", "feature request", "the system should", billing/login problems, or any handoff from support-diagnose after the two-strike rule. Usually reached through support-navigator.

One-time configuration for Cohort Claude Support itself — two minutes, run once after installing the plugin. Records the little that support needs to personalize itself: the member's cohort start date (so "what week am I in" works), their name if the Brain doesn't have it, and a snapshot of what's installed today. Creates support's two memory files (support-log and claude-updates) and confirms the escalation doors from the cohort knowledge base. Reads everything it can from the Brain first and never re-asks what's already known. Trigger on: "set up my support", "set up cohort support", "configure support", "support setup" — or OFFERED at the END of a support session when the member has a Brain but no support config yet. Never as a detour before their actual question is resolved, and never when no Brain exists (the Brain comes first; this is not the first step).