AI agents
Not a prompt. A colleague with a job, a memory, and a lane.
Each agent is a complete Claude Code package: the agent file, its skills, its memory files, an onboarding it runs by chat, and the rituals it keeps. Ten to thirty minutes to install. No API keys required.
The whole teamThe Agent Team
All seven agents in one download: the co-founder that ties them together, both media buyers, the email strategist, the creative director, the app review auditor, and the build foreman. The co-founder reads the other agents' reports and hands you one brief.
$343$149against $343 one at a time
AI agent
The Operator Co-Founder
An AI co-founder that knows your business, your numbers, your non-negotiables, and every lesson you have paid for, and pushes back before you chase the wrong thing.
What is inside- Onboarding by chat: goals, margins, team, working style, non-negotiables, lessons learned, loaded every session
- Engineered pushback: names the weakness, the contradiction, and the shiny object before you commit
- Memory that compounds: decisions, outcomes, and lessons saved automatically in files you own
- Chief of staff for your other agents: reads their reports and hands you one brief
- Rituals: the morning brief, the weekly review, the decision memo, the quarterly reset
- Makes the call, gives the reasoning, waits for your green light. No lists of things to consider
Who it is for- Owners and operators running a real business, any size
- Anyone building an agent team who needs the one that ties it together
- Anyone tired of explaining their business to an AI ten times a week
What you needA Claude account (Pro or Max) and Claude Code. No API keys, no external accounts.
AI agent
The Meta Media Buyer
A media buyer for your Meta account that audits every campaign against your own benchmarks, tells you what to scale, hold, or kill, and logs every move so you never re-run a documented loser.
What is inside- Install audit from your Ads Manager exports: real percentile benchmarks for CPA, ROAS, CTR, CPM, and frequency from your account, not industry averages
- Weekly audit that labels every ad set SCALE, HOLD, FATIGUING, or KILL with the number that decided it
- Budget moves with reasoning and a cap: nothing moves more than the rule you set without your green light
- The winners and losers database, so every proven and failed concept is logged forever
- Test plan builder: what to test next, why, and what result would change your mind
- Optional live connection: the agent writes and runs the Meta Marketing API pulls with your token, and you approve every script before it runs
Who it is for- Owners and marketers running Meta ads who cannot review every ad set every week
- Anyone paying an agency for a Monday audit that arrives on Thursday
- Anyone who has re-tested a losing audience because nobody wrote it down
What you needA Claude account (Pro or Max), Claude Code, and CSV exports from Ads Manager (the agent tells you exactly which). Optional: a Meta Marketing API token for live pulls.
AI agent
The Google Media Buyer
A media buyer for Google Ads that audits search, Performance Max, and YouTube against your own numbers, cleans your search terms, and proposes bids and budgets with the reasoning attached.
What is inside- Install audit from your Google Ads exports: real percentile benchmarks for CPA, ROAS, CTR, and conversion rate from your account
- Weekly audit by campaign and ad group: SCALE, HOLD, FIX, or KILL with the deciding number
- Search terms cleanup: negatives to add, terms to promote to exact match, and the waste they were costing
- Bid and budget moves with a cap, the reasoning, and a rollback
- Landing page and ad copy mismatch flags: where the click and the page disagree
- Optional live connection through the Google Ads API, scripts shown and approved before they run
Who it is for- Owners running search or Performance Max who suspect they are paying for the wrong clicks
- Anyone whose agency sends a report but never a search terms list
- Local and service businesses where one wrong keyword eats the month
What you needA Claude account (Pro or Max), Claude Code, and exports from Google Ads (campaigns, ad groups, search terms). Optional: Google Ads API access for live pulls.
AI agent
The Email Strategist
An email strategist that builds your flows, plans your calendar, writes every send in your voice, and reads your own open, click, and revenue numbers to decide what goes out next.
What is inside- Onboarding that captures your voice, your offers, your customer, and your list segments, loaded every session
- The five flows every business needs, written for yours: welcome, abandoned intent, post-purchase, win-back, and the nurture that sells without pitching
- A monthly calendar with subject lines, angles, and the one call to action per send
- Every send written in your voice, with a subject line test and a plain-text version
- The weekly numbers review from your platform exports: what to repeat, what to retire, the segment that is quietly dying
- Deliverability guardrails: list hygiene rules, send-frequency caps, and the words that get you filtered
Who it is for- Owners with a list and no rhythm
- Anyone paying for an email platform and sending once a month, if that
- Teams whose emails sound like a template because they are
What you needA Claude account (Pro or Max), Claude Code, and exports from your email platform (campaign and flow reports). Works with any platform; nothing to connect.
AI agent
The Creative Director
A creative director that audits your ads against your own winners, briefs the next round with a hypothesis every time, and keeps the winner database so you never re-shoot a documented loser.
What is inside- Install audit from your ad exports: your top and bottom creatives scored on your account's own benchmarks, with the patterns named
- Weekly brief drop: up to ten production-ready briefs across four tracks: AI-made, real customer video, new shoots, and graphics, volume tuned to your spend
- Every brief has a hypothesis, a hook, the visual direction, script beats, the success number, and the proven winner it builds on
- Winner exploitation: five to ten variations of a proven concept before any new concept gets budget
- The winner database: every proven and failed concept logged forever with what it had in common
- Competitor watch from the public ad libraries, when you paste in what you see; stays in its lane: briefs, audits, and creative intel only
Who it is for- Owners and marketers running paid social who cannot keep up with creative volume
- Brands stuck on one style or one creator, watching costs climb
- Anyone paying an agency for four briefs a month
What you needA Claude account (Pro or Max), Claude Code, and ad-level exports with creative names or IDs. Optional: paste competitor ads from the public ad library for the weekly watch.
AI agent
The App Review Auditor
Point it at your project, your screenshots, and your store listing. It audits against Apple's and Google's rulebooks, writes the findings with severity and the guideline behind each, and fixes what you approve.
What is inside- Two rulebooks, Apple and Google Play, as checklists it reads every audit: payments, Sign in with Apple, account deletion, website-only apps, demo login, iPad layout, screenshots, coming soon screens, broken links, report and block, restore purchases, paid features disclosed
- The audit: opens your project, reads the code paths, the screens, the listing text, and the review notes, and produces a findings report with severity and the exact guideline
- The fix plan: for every finding, what to change, where, and the risk. You say go, it makes the change in the project and shows the diff
- Review notes and demo-account writer, screenshot and metadata checker, link checker
- Rejection handler: paste the rejection, it maps it to the finding, fixes it, and drafts the reply
- Memory of every submission, finding, and outcome so the second app is easier than the first
Who it is for- Anyone submitting an app to the App Store or Google Play for the first time, or the fourth
- Builders who used AI to build the app and now need to pass a human review
- Owners paying a developer and wanting an independent check before submission
What you needA Claude account (Pro or Max), Claude Code, and the project folder for your app. Screenshots and listing text as files. No API keys.
AI agent
The Build Foreman
Sets up your machine, connects your tools, writes the loop packet for your app, and runs the build with you at the boundaries. It executes the how; you keep the deploy button.
What is inside- Environment setup it runs with you: checks what is installed, installs what is missing in the right order, for Mac or Windows
- Connector setup: repository, database, payments, browser, with keys in environment variables and a test of each door
- The project brief and the loop packet for your app: laws, definition of done, report format, deploy rules, slice queue
- Runs the build loop: plans slices, builds them, verifies by running or opening, reports PASS or FAIL, stops on two failures
- Model routing built in: strongest for planning, efficient for slices, cheapest for checks
- Ship checklist for both stores and a handoff to the App Review Auditor
Who it is for- Owners and operators building their first real app with AI
- Builders who have a prototype and want it to survive month three
- Anyone who wants the setup done right once instead of guessed at for a week
What you needA Claude account (Pro or Max), Claude Code, admin rights on your machine, and accounts for the services you choose (GitHub, a hosted database, a payment provider). Keys stay in your environment.