Open Source Tools

AI Security Toolkit

Open-source tools to scan, protect, monitor, and comply. Free for individuals and open-source projects.

On the roadmap

Tools we're planning next. Want to help build them or see one prioritized? Get in touch.

Prompt Firewall

Planned
opensecureai/prompt-firewall

Runtime middleware to filter malicious inputs and validate LLM outputs in real-time. Planned support for Python, Node.js, and Go.

Key Features

  • Real-time input filtering
  • Output content validation
  • Custom rule engine
  • PII detection & redaction
  • Drop-in SDK integration

Red Team Toolkit

Planned
opensecureai/redteam

Automated red teaming framework for LLMs. Generate adversarial inputs, test guardrails, and measure model robustness against known attack patterns.

Key Features

  • Automated jailbreak testing
  • Prompt leak detection
  • Data extraction probes
  • Custom attack scenarios
  • Detailed reporting

Model Integrity Checker

Planned
opensecureai/integrity

Verify AI model provenance and integrity across the supply chain. Detect tampering, poisoning, and unauthorized modifications.

Key Features

  • Model hash verification
  • Training data audit
  • Backdoor detection
  • Provenance tracking
  • SBOM generation

Compliance Dashboard

Planned
opensecureai/comply

Map AI deployments against regulatory frameworks. Auto-generate compliance documentation and track remediation progress.

Key Features

  • EU AI Act mapping
  • NIST AI RMF alignment
  • ISO 42001 compliance
  • Risk scoring engine
  • Audit trail generation

LLM Observability

Planned
opensecureai/observe

Monitor LLM applications in production for security anomalies, prompt injection attempts, and unusual behavior patterns.

Key Features

  • Real-time threat monitoring
  • Anomaly detection
  • Usage analytics
  • Alert integration
  • Custom dashboards

Quick Start

Terminal
$ npm install @opensecureai/scanner
$ npx opensecureai scan ./prompts/*.txt
$ npx opensecureai scan prompt.txt --fail-on high