Problem / Scope

AISecurityBase is the public distribution surface for AI security research generated from Tethermark and adjacent review workflows. It is meant to show ranked, reviewed, and clearly bounded AI security findings while building an owned audience through newsletter signup and content sequences.

Architecture

• Leaderboard and result pages for reviewed AI security posture signals
• Ingest boundary for Tethermark-generated scores, finding summaries, and freshness metadata
• Newsletter capture and follow-up path using Beehiiv and Supabase
• Publication workflow that separates private scan execution from public, reviewed research output

Key Workflows / What It Proves

• Turn audit results into explainable public research instead of isolated command-line artifacts
• Build an email audience around AI security findings, benchmark notes, and practical remediation patterns
• Show that the audit harness can support a commercial or media-style product surface
• Create a credible companion asset for outreach, partnerships, and customer discovery

Limitations

• Public rankings should launch only after result normalization, freshness logic, and publication review are tightened
• Beehiiv and email sequence setup must be completed before using the site as the main traffic destination
• The site should support Tethermark's narrative rather than distract from the core harness launch

Evidence Pack

E-ASB-001

Leaderboard and ingest boundary

Local implementation proving a public result surface can consume reviewed AI security posture outputs.

View registry

E-ASB-002

Newsletter capture workflow

Beehiiv and Supabase-backed subscription path intended to convert launch attention into an owned audience.

View registry

E-ASB-003

Launch sequencing note

Recommendation to soft-reference AISecurityBase during early Tethermark posts, then promote it directly once email and result flows are validated.

View registry