About Superagent
Superagent is a security platform made up of AI security workers, each responsible for one job across applications, repositories, agents and infrastructure. It also offers its own decision model for security triage, called Security One, available at a 27B parameter size.
The workers listed on the site are PR Security, which screens pull requests and untrusted contributors before they merge; Secure Dependency Updates, which turns vulnerable dependencies into tested, merge ready pull requests; Red Team, which looks for the exploit paths an attacker would use; and Agent Guardrails, which helps agents run without leaking secrets or breaking policy. Contributor Trust, Agreements and Runtime Guardrails are also included. The documented use cases include scheduled repository red team runs that hand triaged findings to Cursor for remediation, dependency updates validated end to end with Devin, and triage of Supabase logs. Access is available through an API and an MCP server, and existing integrations and workflows stay in place.
It is built for engineering and security teams, including open source maintainers. The vendor is backed by Y Combinator.
Pricing is credit based. Public repositories are free across the whole suite. For private repositories, a Free plan costs $0 with 100 credits a month, Starter is $499 a month for 500 credits, Growth is $1,499 a month for 2,000 credits, and Scale starts at $3,999 a month for 6,000 credits. Annual plans include 20 percent more credits. Every plan includes all workers and unlimited seats. Example rates are 1 credit per PR scan, 5 credits per private dependency update run and 500 credits per Red Team report.