All four are built on one stack: a Ruflo hive + sgit vaults + Agent Behaviour Policies.
1Task runs
2Dedicated compute
3Maintained repository
4Expert time
1Task runson our infrastructure
A brief and your materials go in. A Ruflo queen and her workers, each role under its own behaviour policy, work inside an ephemeral sgit vault with no network. The vault comes back to you.
2Dedicated computeone machine per task
An AWS EC2 instance in London is started for your task and stopped after it. The sgit vault is created inside that machine, so nothing is shared with other customers.
3Maintained repositoryrun it yourself
The full ruv.sgit setup in a private repository of your own, kept up to date by us. Run it on Claude Code, on AWS or wherever you run agents; tokens and cloud stay on your accounts.
4Expert timewith the people who built it
Scope tasks, steer runs, read the vaults with you. Or design a pilot and its Agent Behaviour Policy, or set ruv.sgit up in your environment and hand it over.
+Mix and matchthe four combine
Consultants
White label it
Put your name on it. Your client receives the run's vault.
Companies
Add it to your offering
From reselling runs to your own deployment. Your clients, your brand, your keys.
Internal teams
Provide it inside
Agent swarms for your organisation, with the governance already written.
Anybody who implements this implements sgit vaults and Agent Behaviour Policies.
Anybody who implements this implements sgit vaults and Agent Behaviour Policies.
ruv.sgit.ai/businessagent@riskmandate.ai
Prices in GBP, excluding VAT, from data/offers.json. Built on Ruflo (MIT, by rUv / ruvnet); sgit vaults and ABPs are ours (Apache-2.0).