In biology
Genes that get transcribed as one unit. They belong together.
About
Make your whole team as good at AI as your best person.
The best AI prompts on your team live on individual laptops. Operon finds the ones that recur, polishes them, and ships them as agents your whole team can install.
01
Why the name "operon"?
An operon is a unit of biology. A cluster of genes that get transcribed together as one functional bundle. That's a useful prompt: not one line, but a cluster of steps that fire as a unit.
In your team
The steps your best engineer takes when they say "review this PR." Captured once, runnable by anyone.
As a product
A versioned, installable agent anyone on your team can run with one command. Same input, same result.
02
Why we built it.
Every team has one engineer who figured out which prompts actually work. That knowledge lives on one laptop.
Stuck on one laptop
Good prompts live in CLI history and people's heads. Nobody else gets to use them.
Lost to Slack
The patterns that work get DM'd around and lost. Search Slack for "claude prompt" and weep.
Walks out the door
When your best AI user leaves, six months of prompt craft leaves with them. No handoff. No backup.
03
How it works.
Operon plugs into Claude Code, Codex, and Cursor. No new habits. Local‑first scan, redaction on the laptop. You ship only what the team chooses to publish.
1
Mine
Scan local sessions for recurring prompts. Nothing leaves the laptop until you say so.
$ operon mine --since 30d 2
Polish
Claude drafts a clean SKILL.md from each candidate. You review, edit, and publish.
$ operon polish review-pr 3
Share
One command installs the operon into any teammate's Claude Code, Codex, and Cursor.
$ operon install acme/review-pr