Skip to content

[security][research] Evaluate privacy-preserving multi-signal Sybil correlation #90

Description

@RobLe3

Research-only scope

Evaluate whether future Sybil correlation can use multiple network-provenance signals (source IP/prefix, ASN/provider class, registration timing, operator timing, reciprocal credits, repeated capability fingerprints) as heuristic evidence. This must not be treated as identity proof: shared NAT/CGNAT, universities, enterprises, and cloud providers can legitimately contain many operators.

Before collecting or retaining any new signal, document purpose limitation, data minimization, retention/deletion, operator notice, GDPR/EU AI Act implications, false-positive handling, appeal paths, and whether the signal can be derived from existing ephemeral evidence. Do not add ASN/IP correlation to scoring, badges, federation trust, or public output from this issue alone.

Deliverable

A threat-model and privacy decision record that classifies each signal as accepted, rejected, or deferred; defines content-free aggregate evidence; and requires multi-signal correlation plus human/operator review for any future mitigation. Add negative tests for legitimate shared-network cases. Implementation is explicitly deferred pending this decision.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    proposalProtocol or conformance proposal

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions