Security and privacy

Bounded threat model

This model describes the static product boundary and its residual risks. It is not a security certification and does not claim server-side privacy outside the implemented Netlify and third-party boundaries.

Browser-Local Saved Work

Threat: Another person or extension with access to the browser profile can inspect local data.

Controls: Keep data browser-local; explain shared-device risk; provide deletion and export controls.

Residual risk: Control reduces but does not eliminate risk.

Finder Selections

Threat: Selections could reveal civic interests if placed in URLs or telemetry.

Controls: Keep selections out of URLs and analytics; use local state only.

Residual risk: Control reduces but does not eliminate risk.

Feedback/Correction Submissions

Threat: A submission can contain voluntary identity or sensitive claims.

Controls: Minimize fields, make contact optional, require transmission notice, keep form output non-public.

Residual risk: Control reduces but does not eliminate risk.

Analytics

Threat: Fine-grained paths or content events could create civic-interest profiles.

Controls: Require affirmative consent and send only broad route family plus viewport class.

Residual risk: Control reduces but does not eliminate risk.

Support/Payment Boundary

Threat: Third-party support or payment services have separate data practices.

Controls: Label outbound boundaries; do not imply WS Musings controls the third party.

Residual risk: Control reduces but does not eliminate risk.

Static Apis

Threat: Machine records can be misread as verified findings.

Controls: Use route/indexation controls, readiness terms, provenance, and interpretation boundaries.

Residual risk: Control reduces but does not eliminate risk.

Supply-Chain/Build Integrity

Threat: A compromised dependency or build input can alter public output.

Controls: Pin governed inputs, hash releases, validate deterministically, and perform recovery drills.

Residual risk: Control reduces but does not eliminate risk.

Malicious Or Malformed Imports

Threat: Untrusted imports can inject markup or corrupt governed records.

Controls: Validate schemas, escape public text, reject unknown fields, and fail closed.

Residual risk: Control reduces but does not eliminate risk.

Source-Link Abuse

Threat: A linked authority can redirect, disappear, or become malicious.

Controls: Track resolution, status, stale threshold, owner, and next review; propagate broken state.

Residual risk: Control reduces but does not eliminate risk.