CVP CALENDAR

Anthropic Cyber Verification Program — Sunglasses public run log

In April 2026, Anthropic approved Sunglasses for their Cyber Verification Program (CVP) — a narrow lane for responsible cybersecurity evaluation of frontier models.

This calendar is our public log. Every published run gets its own dated report. Days without a public run carry an internal-testing note so you can see the cadence is honest.

Current cadence: two runs per week, each with fresh threat-class prompt sets. Approved: April 16, 2026. Run 1: April 17, 2026.

Published run

Scheduled

Internal testing

Today

April 2026

What happens on each run

Each run follows the same methodology:

3 prompts frozen before execution (benign defensive / borderline legitimate / clearly high-risk)
model runs in an isolated Claude Code session on the CVP-approved org
full transcripts captured, hashed, and preserved
each response scored on class (allowed / partial / blocked), usefulness, and safety
evidence bundle archived internally; public report published here

Each report includes its methodology, per-prompt scoring, limitations, and honest nuance. We publish what the model did well and where it fell short.

Why we publish this

Many teams talk about AI safety. Fewer publish evidence. Approval into Anthropic's CVP gave us a narrow, authorized path to evaluate frontier model behavior — we think the honest move is to show the work in public.

We are not trying to prove Claude "safe" or "unsafe." We are mapping where useful-for-defenders ends and where blocked-for-misuse begins, one run at a time.

Free. Open source. Honest evidence bundles.

sunglasses.dev
github.com/sunglasses-dev/sunglasses

Anthropic's Cyber Verification Program approved Sunglasses on April 16, 2026. Reports on this page are produced under that program.
Not affiliated with Anthropic. SUNGLASSES is a free, open-source project.