Tell your computer what you want it to become.
Linura turns human and agent intent into policy-controlled, verified machine state — across Linux hosts, AI agents, containers, and virtual machines.
Website · Core repository · Docs · Discussions · Releases
Linura is an intent-driven system layer for Linux. Humans and AI agents can describe what a machine should become; Linura keeps the consequences behind a deterministic authority boundary that plans, authorizes, executes, verifies, audits, and reconciles managed change.
The intelligence layer may be probabilistic. Machine authority is not.
human / agent intent
↓
propose
↓
observe → plan → validate → authorize → prepare
↓
execute → verify → commit → audit → reconcile
↓
verified machine state
Agents propose. Linura decides and executes.
| Start with the project | linura-org/linura — architecture, implementation, release evidence, and canonical project policy |
| Understand the architecture | Architecture — the control model, trust boundaries, and authority lifecycle |
| See where it's going | Roadmap — current milestones and development direction |
| Contribute | Contributing guide — first contributions through architecture/security-sensitive work |
| Join the conversation | Discussions — Q&A, architecture, RFCs, ideas, and show-and-tell |
| Report actionable work | Issues |
| Report a vulnerability privately | Security policy |
| Use verified artifacts | Releases |
Linura is currently experimental and pre-1.0. Supported-platform boundaries, qualification evidence, compatibility claims, release status, governance, and security policy are version-scoped and remain canonical in the linura repository.
This organization profile is intentionally a front door, not a second source of truth.