# What Airship secures

> The targets Airship's agents cover, from web apps and AI agents to binaries, kernels and smart contracts.

Airship secures the whole stack, not one layer of it. Each target below gets context-aware analysis, because Airship's agents read your code, business logic and architecture instead of matching known signatures.

## Web apps and APIs

Agents probe REST, GraphQL and web apps the way an attacker would. Airship discovers endpoints, exercises them, and attaches a proof-of-exploit to every finding.

## Agent systems

LLM endpoints, MCP servers and stateful AI agents. Airship tests how they handle untrusted input, what their tools can reach, and whether an attacker can steer them.

## Source code and repositories

Git repositories, application code and C/C++ libraries. Airship analyzes your code and pull requests for security issues, so vulnerabilities are caught at the source. See [CI/CD and pull requests](/docs/ci).

## Software packages and dependencies

npm, PyPI and Cargo packages. Airship checks the packages you publish and the ones you depend on.

## Containers and images

OCI container images, inspected layer by layer for vulnerable software and risky configuration.

## Compiled binaries

Native executables, compiled libraries and other binary artifacts, analyzed with or without source access.

## Operating systems and kernels

The Linux kernel, Apple XNU, kernel extensions and Windows Hyper-V research.

## Smart contracts and Web3

EVM, Solana, Cardano, Starknet/Cairo and Move contracts.

## Infrastructure and cloud

Airship looks for misconfigurations and exposures across your infrastructure and cloud environments, before an attacker finds them.

## Continuous coverage

Security isn't a one-off at deploy time. Airship monitors your attack surface around the clock and tests new CVEs against your systems automatically.

> [!IMPORTANT]
> Only test systems you own or have written permission to test. See [Authorized testing](/docs/authorized-testing).