Rogue

Free plan Self-hosted

What Rogue is

Rogue is a self-hosted platform for evaluating AI agents and examining their security. Its automated evaluations compare an agent’s behavior with business policies and expected behavior, then produce pass/fail reports with reasoning. For security work, Rogue covers more than 75 vulnerabilities across 12 categories and 20 attack techniques, including encoding, social engineering, injection, semantic, and technical attacks. Vulnerabilities receive CVSS-based risk scores from 0 to 10.

Teams can map testing to frameworks including OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF, ISO/IEC 42001, the EU AI Act, GDPR, and OWASP API Top 10. Rogue provides a server, terminal interface, and non-interactive CLI for CI/CD pipelines. It supports A2A agents over HTTP, MCP agents over SSE or Streamable HTTP, and direct Python calls. OpenAI, Anthropic, and Google models are supported through LiteLLM, and getting started requires an LLM API key. Reports can be exported as Markdown, CSV, or JSON, and scans can use a random seed for reproducibility. Basic scans use five curated vulnerabilities and six attacks; full scans use more than 75 vulnerabilities and 40 or more attacks.

Who it suits

Rogue suits security teams and developers building or deploying AI agents who need automated evaluations, regression testing, security audits, or compliance reporting. Its self-hosted deployment and CI/CD interface also fit teams running agent checks within their own systems.

What it does well

Automated pass/fail evaluations include reasoning.
Covers 75+ vulnerabilities across 12 security categories.
Maps testing to frameworks including NIST AI RMF and GDPR.
Supports A2A, MCP, and direct Python agent protocols.
Exports reports as Markdown, CSV, and JSON.
Random seeds can make scan results reproducible.

What to know first

Getting started requires an LLM API key.
Personal and internal use is free; commercial hosting requires licensing.
Basic scans cover five vulnerabilities and six attacks.

Limits

The free plan is for personal and internal use; commercial hosting requires licensing. Basic scans use five curated vulnerabilities and six attacks, while full scans use more than 75 vulnerabilities and 40 or more attacks.

Plans

Plan	Billed and limits	Price
Personal and internal use	Free for personal and internal use	Free
Commercial hosting	Requires licensing · Contact [email protected]	On request

As github.com lists them, read 1 Oct 2026. The maker's own page has the final word.

Getting started

Open the Rogue repository on GitHub.
Prepare uvx and Python 3.10 or newer.
Get an API key from OpenAI, Anthropic, or Google.
Run Rogue using its documented quick start.
Choose the server, terminal interface, or non-interactive CLI for your workflow.

Verdict

Choose Rogue if you need a self-hosted way to evaluate agent behavior and run security scans with risk scores and compliance mappings. Look elsewhere if you need commercial hosting without licensing or want broader coverage than the basic scan provides.

Questions

How much does Rogue cost?

The personal and internal use plan is free. Commercial hosting requires licensing, with pricing on request.

Where can Rogue run?

Rogue is self-hosted. It provides a server, terminal interface, and non-interactive CLI.

Which AI models does Rogue support?

It lists OpenAI, Anthropic, and Google models through LiteLLM. An LLM API key is required to get started.

Which agent protocols are supported?

Rogue lists A2A over HTTP, MCP over SSE or Streamable HTTP, and direct Python function calls.

What does the license allow?

The Qualifire OSS License combines MIT terms with a Commons Clause that excludes selling the software or providing it as a paid hosted service. Rogue is free for personal and internal use; commercial hosting requires licensing.

What reports can Rogue export?

Reports can be exported in Markdown, CSV, and JSON formats.

Facts from Rogue's own pages

Commercial use: The README says Rogue is free for personal and internal use and that commercial hosting requires licensing; it provides [email protected] for contact. · github.com, 30 Sept 2026
Compliance: Rogue maps testing to eight compliance frameworks including OWASP, MITRE, NIST, GDPR and the EU AI Act. · github.com, 1 Oct 2026
Coverage: The repository states that Rogue covers 75+ vulnerabilities across 12 security categories and 20 attack techniques. · github.com, 1 Oct 2026
Evaluation: Automatic Evaluation tests agents against business policies and expected behaviors with pass/fail reports and reasoning. · github.com, 1 Oct 2026
Frameworks: The documented framework coverage includes OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF, ISO/IEC 42001, EU AI Act, GDPR, and OWASP API Top 10. · github.com, 30 Sept 2026
Interfaces: Rogue provides a server, terminal user interface and non-interactive CLI for CI/CD pipelines. · github.com, 1 Oct 2026
License: The repository license is the Qualifire OSS License, combining MIT terms with a Commons Clause that excludes selling the software or offering it as a paid hosted service. · github.com, 30 Sept 2026
Maker: The license identifies Qualifire ltd as Rogue's licensor. · github.com, 30 Sept 2026
Model providers: Rogue lists OpenAI, Anthropic, and Google models via LiteLLM and requires an LLM API key to get started. · github.com, 30 Sept 2026
Models: Rogue supports OpenAI, Anthropic and Google models through LiteLLM. · github.com, 1 Oct 2026
Product: Rogue is an AI agent evaluation and red teaming platform for testing agent reliability and security. · github.com, 30 Sept 2026
Protocols: Supported agent protocols are A2A over HTTP, MCP over SSE or streamable HTTP, and direct Python function calls. · github.com, 1 Oct 2026
Purpose: Rogue is an AI agent evaluator and red team platform for stress-testing agents before attackers do. · github.com, 1 Oct 2026
Red teaming: Red Teaming simulates adversarial attacks to find security vulnerabilities. · github.com, 1 Oct 2026
Reports: Rogue exports comprehensive reports in Markdown, CSV and JSON formats. · github.com, 1 Oct 2026
Reproducibility: Scans can use a random seed to make results reproducible. · github.com, 30 Sept 2026
Requirements: The quick start lists uvx, Python 3.10+, and an API key from OpenAI, Anthropic, or Google as prerequisites. · github.com, 30 Sept 2026
Risk scoring: Rogue assigns vulnerabilities a CVSS-based risk score from 0 to 10. · github.com, 30 Sept 2026
Runtime integrations: The AI AppSec product page lists LangChain, CrewAI, AutoGen, custom builds, SIEM systems and webhooks, including Splunk, Datadog and PagerDuty. · rogue.security, 1 Oct 2026
Scan limits: Basic scans use 5 curated vulnerabilities and 6 attacks, while full scans use 75+ vulnerabilities and 40+ attacks. · github.com, 1 Oct 2026
Security posture: Rogue Security states that its Trust Center provides information about SOC 2 compliance and data handling, and that it supports end-to-end encryption and zero-data-egress in-VPC deployment. · rogue.security, 1 Oct 2026
Target users: The product is presented for security teams and developers building or deploying AI agents, including teams needing regression testing, security audits and compliance reporting. · github.com, 1 Oct 2026

Where Rogue stands on Freedom251

AI Red Teaming Tools#17 of 27