Security insights and product updates.

Last week I wrote about building a scanner that writes its own exploits. A cron job runs Claude Code every night, it reads new CVEs, writes proof-of-concept checks, tests them against vulnerable and patched Docker instances, and opens PRs. I review and merge in the morning. That post was about how the pipeline works. This one is about what happens when you point it at 1,500 real servers. The target list I pulled 1,500 IPs from Shodan tagged as GitLab instances. Only 41% were actually runn

Every night at 3AM, a cron job on my homelab server runs claude /cve-pipeline. By morning, there are pull requests waiting. Each one is a working exploit check for a recent CVE, tested against a vulnerable instance, tested again against a patched instance, with WAF bypass variants for the most popular WAF setups. In the morning I review the PRs, merge, deploy. Users get the new checks on their next scan. I'm a solo founder. The whole thing runs on a $200/mo Claude Max subscription. 31 expl

CC7.1 never names a scan frequency or a tool. Auditors still expect scan output on a stated cadence, findings fixed within your SLAs, and a scope that matches your system description. Here is the exact criteria text, what Type I and Type II need, and where an external scanner fits.
58% of 3,369 startup domains had a failing security check and 4% had a high or critical finding.