Harbor Social · Public

Harbor Social Platform Integrity

Impersonation, malicious-link and AI-generated abuse detection across Harbor's sandbox graph.

Authorized testing only

—Use only approved accounts, systems, and scenarios.
—Do not interact with uninvolved people.
—Do not cause real financial loss.
—Do not access accounts without authorization.
—Do not obtain or submit stolen credentials.
—Do not harass, dox, threaten, or extort anyone.
—Stop testing if you encounter real victim harm.

Authorized scope

Sandbox content graphSynthetic posts and profiles

Link safety pipelineApproved benign test domains only

In-scope weakness categories

S1WE-300

Trust & Reputation

Manipulation of reputation, trust, tenure or legitimacy signals.

S1WE-500

Content & Discovery

Listings, ads, profiles, recommendations, malicious content and distribution.

S1WE-900

Coordinated Abuse

Graph, cluster, cross-account, cross-device and campaign-level detection.

S1WE-1000

Automation & AI Abuse

Bots, automated abuse, AI agents and synthetic abusive content.

Public-safe disclosures

S1AV-2027-004516.4MediumResolved

Enforcement state inconsistent across approved test-account recovery flow

A restricted sponsor test account regained a subset of capabilities through the recovery flow, leaving enforcement state inconsistent between services.