XERXES SI / ARCHITECTURE / ALPHA

SysOp Rootless Daemon Roadmap: Operating Agents Without Granting Root

Operations and security are different responsibilities. Their daemons should remain separate from one another and from XMAN’s operator authority.

Reviewed against XMAN 0.4.76a1 source · November 1, 2026 · No independent certification claimed

PLANNED COMPONENT

The planned SysOp daemon

The SysOp is envisioned as an operations specialist with its own rootless daemon, narrowly scoped to health checks, status inspection and proposed maintenance actions. The SysOp would not acquire XMAN operator credentials merely because it identifies a problem. Operator approval and XMAN policy still decide whether a privileged host action is authorized.

Status: roadmap only. No completed daemon API, release date, production monitoring coverage, automated repair success rate or independent security qualification is claimed.

ROLE SEPARATION

Independent roles, separate trust decisions

XMAN manager daemon: current Alpha

Owns bounded resource admission, durable requests, scheduler decisions, scoped approvals and the limited Podman adapter.

SysOp daemon: proposed

Owns a limited operations and health-observation scope under separate credentials and service lifecycle, with no unilateral host privilege escalation.

Security Officer daemon: proposed

Owns a separate future security analysis scope, such as plugin preflight checks and runtime incident evidence; potential containment requires explicit grants and safety tests.

ENGINEERING REQUIREMENTS

How the services should communicate

All integrations should use a documented API contract with authenticated identity, narrowly scoped methods, idempotency, generation-bound targets, explicit authorization and correlated receipts. Events should be treated as untrusted observations until independently verified. No model-generated string can be construed as a human approval.

The SysOp’s repair proposal should be inspectable before execution; the Security Officer’s incident reports should distinguish suspicion, authorization, action attempted and verified outcome. A false positive should have an operator-controlled recovery path.

TRANSPARENT LIMITS

What this is not

It is not a claim that Clawsolator already prevents every prompt injection or malicious plugin, and it is not a general-purpose OS-root watchdog. The future daemons must be designed, threat-modeled, tested and independently qualified. We respect the OSSTMM testing methodology without claiming certification.