> For clean Markdown of this page, append .md to its URL. For the complete documentation index, see https://www.superagent.sh/llms.txt.


Protect deployed applications with adversarial testing, context controls, finding handoffs, and retained results.

# Applications

Applications are deployed software with public or internal interfaces, data
flows, identities, and release controls.

## Available products

| Product | What it does for applications |
| --- | --- |
| [Red Teaming](https://www.superagent.sh/docs/products/red-teaming/applications) | Tests live applications through a real browser and records reproducible attack paths |
| [Context Guardrails](https://www.superagent.sh/docs/products/context-guardrails) | Checks external content sent to agents |
| [Runtime Guardrails](https://www.superagent.sh/docs/products/runtime-guardrails) | Monitors or blocks risky agent actions |

## Add an application

1. Open [Applications](https://www.superagent.sh/app/applications) and select **Add new**.
2. Enter a name and the deployed application's URL.
3. Optionally select a connected repository to provide source context for
   later work.
4. Select **Create application**.
5. Open the application to configure Red Teaming access, credentials, request
   limits, goals, and schedules.

### Import from Cloudflare

Only organization owners can import and approve Cloudflare hosts.

1. If Cloudflare is not connected, connect it in [Integrations](https://www.superagent.sh/app/integrations) with a user API token that has **Zone:Read** and **DNS:Read**.
2. In the application's **Inventory** tab, select **Import → Cloudflare**, then pick an account and the hosts to import.
3. Mark approved hosts **Included** before running a red-team assessment. New hosts start as **Unverified**. Existing hosts keep their scope settings.

Turn on **Keep inventory in sync** to refresh hosts daily. Review new hosts in **Inventory** and change which zones sync under **Import → Manage sources**.

## Close the loop

Route verified findings to the customer agent or workflow that owns the application. The agent returns a structured agent report through Superagent MCP. Superagent retains that context and applies it to later application, repository, agent, and infrastructure work.

## Human control

Keep deployment, disclosure, release, and production-change approval with the responsible people and systems.

## Next steps

- [Run application Red Teaming](https://www.superagent.sh/docs/products/red-teaming/applications)
- [Check agent inputs with Context Guardrails](https://www.superagent.sh/docs/products/context-guardrails)
- [Control agent actions with Runtime Guardrails](https://www.superagent.sh/docs/products/runtime-guardrails)

---
Source: https://www.superagent.sh/docs/concepts/applications
Index: https://www.superagent.sh/llms.txt
