Superagent

What it does

Superagent is a security operations platform for teams protecting software they build and run, including teams using coding agents. Connect a repository, application, or agent, select a security feature, then review the resulting check, report, context result, or runtime decision in the dashboard. It supports PR Security reviews of code and dependency changes, policy-checked dependency update pull requests, adversarial Red Team assessments, and agent guardrails.

The service is intended for software and security teams that want to assess software or control agent interactions as part of development. Repository and application coverage is documented; infrastructure protection is described as coming soon. Setup starts by creating an account and organization, connecting a target, and choosing a feature. GitHub is documented as a repository connection; MCP, webhooks, and a REST API are optional ways to connect integrations or automation. People remain responsible for consequential merge, deployment, and policy decisions.

Pricing uses monthly credits for private repositories, with a free tier and paid plans; annual paid plans are billed upfront and include 20% more credits. Public open-source repositories have unlimited usage at no charge. Usage varies by feature, including per-scan, per-run, per-report, and per-context-item charges; Security-One 27 B is metered by input tokens. The dashboard is the documented starting point; API and MCP access are included in every plan.

What it can do

• Review pull request code, workflow, dependency, and contributor risk
• Generate policy-checked dependency update pull requests
• Run adversarial Red Team assessments and identify attack paths
• Check external content before an agent consumes it
• Monitor or block risky agent actions at runtime
• Route findings into development workflows and retain structured results

Works with

• GitHub
• MCP
• Webhooks
• REST API
• Cursor
• Claude Code
• Codex

Before you choose

• Infrastructure protection is listed as coming soon; current documented target coverage includes repositories, applications, and agents.
• Repository and application assessments require connecting the relevant target and granting access to the surface being tested or monitored.
• Pricing depends on monthly credits and feature-specific usage; annual paid plans are billed upfront, not monthly.

Listed under

• Security
• App and code security
• AI agents
• Appsec
• Dependency updates
• Guardrails
• Open source
• Red teaming

Get started

Create or select an organization in the Superagent dashboard.

1. Create an account and create or select an organization in the dashboard.
2. Connect a software target, such as a GitHub repository, and confirm Superagent can access the surface to assess or monitor.
3. Choose one feature, configure it using its documentation, run the first job, and review its result.

Private repository plans use monthly credits, which refresh monthly. Annual paid plans are billed upfront and include 20% more credits. Open-source repositories receive unlimited usage across the suite. Usage rates include 1 credit per PR Security scan, 5 credits per private Secure Dependency Updates run (public and dry runs are free), 500 credits per Red Team report, and 1 credit per context item checked. Security-One 27 B is metered at 0.05 credits per 1 million input tokens. Contributor Trust, Agreements, and Runtime Guardrails are included.

Source: superagent.sh/pricing, checked 1 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• Superagent | The autonomous security factory ↗www.superagent.sh
• Features | Superagent Docs ↗www.superagent.sh
• Start here | Superagent Docs ↗www.superagent.sh
• Flexible AI security pricing plans | Superagent ↗www.superagent.sh
• About - Superagent ↗www.superagent.sh
• Overview | Superagent Docs ↗www.superagent.sh

Company website

superagent.sh: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.