LilRAE is a local purple-team lab for human operators and AI agents.
It runs intentionally vulnerable services for authorized research, training, and assessment.
Scenario documents select the range. The lilrae CLI manages its lifecycle and run records.
Accepted to Black Hat USA Arsenal 2026, SecTor Arsenal 2026, and SecTor 2026 Briefings.
Active development. Not for production. Not hardened.
Use a dedicated, rebuildable environment and keep unrelated credentials elsewhere. Operate only on systems you are authorized to test. Native Docker can expose the host. A VM boundary reduces risk but does not remove it.
Seats have no default outbound network controls. Read the execution boundaries and limits before starting a lab.
Follow Run your first lab. For an older installation, use the package migration guide.
Check the prerequisites and host requirements. These pages own the supported versions and resource limits.
Use the CLI, MCP, and web references for supported interfaces.
Use lilrae lab info for the current project's URLs and credential locations.
The installed LilRAE startup catalog determines available scenarios. Scenario Pack installation describes selection and acquisition. Reusable definitions live in OpenRAE/env-packs. LilRAE owns local admission, realization, readiness, and operation.
The published manual has four reader routes:
| Reader need | Route |
|---|---|
| Learn through an exercise | Tutorials: first lab |
| Complete a task | How-to guides: install LilRAE |
| Look up exact facts | Reference: CLI |
| Understand reasons and relationships | Explanation: architecture |
- Get support.
- Contribute.
- Report a vulnerability privately.
- Read the security policy and OpenSSF assessment.
Follow applicable laws and obtain authorization before testing. Do not disclose vulnerabilities in public issues or discussions. Lab test credentials are fixtures. Keep generated control-plane credentials private. LilRAE uses standard security tools and gives agents Kali access. This public repository adds no red-team enhancements to their existing capabilities.
MIT.