Product marketing opsfor AI agents
Open-source skills for the operations side of B2B SaaS product marketing: run launches Product, Sales, and Marketing agree on, build sales content reps actually use, and prove what changed. Tested on real launches. For Claude Code, the Claude app, Codex, Cursor, and any agent that reads skills.
Agree. Build. Prove.
Run pmm-ops-context once. It writes down your product, buyers, how it's sold, the leaders who sign off, launch tiers, and where claims and sales content live. Then the three skills run in order: agree on the story before anything is built, build what reps use from that story, then prove what changed.
Where one skill hands work to another
What every skill refuses to do
The same four rules run through the whole library, so a launch story, a battlecard, or an impact readout holds up when someone checks it.
Agree on the target before anything is built.
Nothing gets measured or shipped that wasn't planned first.
Every claim has proof and a product owner.
Someone confirmed it's true today. Nobody writes claims from memory.
Unknowns stay open.
Each one gets an owner and a date instead of a guess.
A link is reported as a link.
Not as proof. And a shared number is never PMM's alone.
Start from the question you're asked
Every piece answers a question someone actually asks a product marketer. Find yours, and it opens the piece the agent will use.