A CMO's marketing OS, with the files.
Every system below runs my real work each week. What you get is the public version: same structure, prompts and rules, with company data removed.
Knowledge base
Re-explaining context in every chat
- What it does
- One page per fact, linked, with sources. The agents read it before they answer any work question, so answers come from my notes and not from the model's guesses.
- What you get
- Folder template, page schema, ingest policy and a linter that flags broken links and missing sources.
Research agent
Reading every source by hand
- What it does
- Plans the question, fetches every source, refuses stats it can't quote, and fact-checks before it saves an HTML report.
- What you get
- The
research skill and its citation rules.
Weekly routines
Putting together the Monday report
- What it does
- The marketing report, a Saturday knowledge scan and a monthly reconcile run on a schedule and post a short summary.
- What you get
- Routine specs as patterns, without data.
Originality check
Guessing if a draft sounds generic
- What it does
- Scores how templated a draft reads, so I can tell when it sounds like the default AI answer.
- What you get
- The
content-originality skill.
Marketing OSINT
Digging through a competitor's site
- What it does
- A dossier on any company from public data: positioning, channels, content, what customers and AI assistants say.
- What you get
- The
marketing-osint skill, without the company lens.
Small tools
Checking LinkedIn profiles one by one
- What it does
- Decision support, UX microcopy and a LinkedIn watch list.
- What you get
- Three skills.
Confirm each system is your own build before publishing; installed third-party packs must not appear here. Keep only the crossed-out lines that are true for your own work.
Install
One repository, one command.
# clone the public OS
git clone repo URL
Format not decided: Claude Code skills, an MCP server, or both.