Catalog

The bot almost every desk on X starts with. One owner that assigns, checks, and comes back only when a human has to decide.

The GTM bot SpaceXAI operators keep iterating: score intent, draft in the seller's voice, never send unattended.

The God-mode pattern from X: many scouts, a tight ranker, and a final pass that kills unverifiable claims.

Focused research, not a firehose. Built from the one-job rule people learned in the first week of Grok Bot.

The builder pattern people are posting this week: CoS instructs, this bot files issues and lands PRs.

A 7am inbox summarizer people stood up in minutes. It does not answer mail. It tells you what is actually due.

The launch-day sales bot: an inbox of personalized drafts waiting for a human to hit send.

The meta-bot. It watches the rest of the army so you do not learn failure modes the hard way.

A playtester that lives on your phone and talks to Cursor. It finds stalls in the game, not in the chat.

Straight from the SpaceXAI launch examples. Hygiene and commit risk, not another forecast speech.

A loop tester with a cloud computer. It clicks through the product, captures breakage, and files only what it can reproduce.

A brief that lands before you walk in. Who you are meeting, what you last discussed, three questions worth asking.

The personal ops bot people actually keep. It watches dates and packages, not your whole life.

A single-job support bot. It watches refunds and slipping dates, then writes the escalation — it does not send it.

A relocation-minded scout. It watches a company list and only surfaces roles that actually match the constraints.

An idea inbox fed by X and LinkedIn, not a ghostwriter. It shows receipts before it suggests a take.

Post-call closer from the launch notes. The work lands in the tool a human would have used.

The bot that keeps the demo from dying in front of a customer. It checks the product UI, not a status page.

A specialist a Chief of Staff can hire. It knows the freight numbers so the rest of the team does not guess.

A skeptical forecast. It argues with the commit and shows the evidence gap.

A GTM specialist for the last mile: slides that match the team's format, not a new design language.

A scheduler that defends focus instead of filling holes. It proposes times; it does not accept for you.

A quiet ledger. It does not lecture about budgets. It shows the week and the oddities.

find relevant conversations, communities and opportunities for things we have already created.

check pricing, positioning, offers, landing pages and launches, then flag meaningful changes.

go through calls, interviews and feedback to find recurring pains, objections, buying triggers and language.

research search, X, competitors and conversations before deciding what is actually worth creating.

take an approved idea and find evidence, examples, data, counterarguments and current conversations around it.

pull everything together and surface the few things that actually deserve my attention.

monitor important queries and pages for ranking, intent and competitor changes.

research qualified inbound accounts and prepare useful context before I speak to them.

monitor performance and investigate unusual movements instead of waiting for me to check another dashboard.

keep hypotheses, variants, results and learnings organized so we do not keep relearning the same things.

Run the job described in this public setup

confidence, impact ranges, base / bull / bear.

CRITICAL / HIGH / MEDIUM with one-sentence impact, why it moves, estimated range, invalidation, sources.

sentiment, technicals, catalysts, options overlay, portfolio impact.

Watch named tickers and alert only on material moves