Trust

Make the evidence boundary visible.

Trust should come from inspectable mechanisms and bounded claims, not from a broad “secure” or “AI-powered” label.

Exact identity

Bind evidence to source/archive/configuration fingerprints and invalidate it when decision-critical inputs change.

No false green

Unavailable native, device, store, network or deployment evidence stays pending—not PASS.

Fail narrower

Missing/conflicting authority should reduce capability or preserve uncertainty rather than silently expand permission.

Public claim discipline

Do not advertise a store, price, certification, provider capability or platform behavior until the corresponding evidence exists.

Separate user decisions

Advisory analysis should remain distinguishable from source facts and user-owned final decisions.

Provider transparency

External stores, AI providers and data services remain separate trust/privacy boundaries under their own terms and availability.

How to inspect RuleOak's open-source claims

The v0.50.0 project publishes source manifests, reproducibility checks, compatibility boundaries, release-verification guidance and security/non-goal documentation. These improve auditability without pretending to be independent certification.

Claims should be evidence-scoped.

Source checks, CI, packaging, native execution, deployment and independent review prove different things.

See the claims and evidence matrix →