llms.txt@packages/website/public git:20260322.d4063b2A

Immutable. This exact content never changes and is served at /api/v1/blob/dcc24551ff4f2863.

# Panguard AI

> Panguard AI is a 100% free, open-source AI agent security platform.
> Skill Auditor, Guard, Scan, MCP Server, ATR rules, and Threat Cloud --
> all MIT licensed. One curl command to install. Zero paywalls.

## Core Products

- [Product Overview](https://panguard.ai/product): The security flywheel -- audit, guard, share, strengthen
- [Skill Auditor](https://panguard.ai/product/skill-auditor): 8-layer pre-install security gate for AI skills
- [Panguard Scan](https://panguard.ai/product/scan): 60-second security audit with Sigma, YARA, and ATR rules
- [Panguard Guard](https://panguard.ai/product/guard): 4-agent AI pipeline -- detect, analyze, respond, report -- 24/7
- [MCP Server](https://panguard.ai/product/mcp): 11 panguard_* tools for Claude Code, Cursor, and AI assistants
- [ATR Standard](https://panguard.ai/atr): 61 open-source rules for AI agent threats (like Sigma for agents)
- [Threat Cloud](https://panguard.ai/threat-cloud): Collective intelligence -- auto-generates rules from real attacks

## Coming Soon

- [Panguard Trap](https://panguard.ai/product/trap): Intelligent honeypot system (under development)
- [Panguard Report](https://panguard.ai/product/report): AI compliance reports for ISO 27001, SOC 2, TCSA (under development)

## License

- 100% free and open source. MIT License. All features included. No paywalls.

## Technology

- [How It Works](https://panguard.ai/how-it-works): 3 detection layers, confidence-based auto-response
- [Threat Cloud](https://panguard.ai/threat-cloud): Collective threat intelligence with 8,000+ detection rules (Sigma + YARA + ATR)

## Documentation

- [Getting Started](https://panguard.ai/docs/getting-started): Install Panguard in 60 seconds
- [CLI Reference](https://panguard.ai/docs/cli): 23 commands
- [Skill Auditor Docs](https://panguard.ai/docs/skill-auditor): AI agent skill security auditing
- [Guard Docs](https://panguard.ai/docs/guard): 4-agent pipeline, 11 response actions, confidence thresholds
- [Scan Docs](https://panguard.ai/docs/scan): Scanner modules and result interpretation

## Agent Threat Rules (ATR)

- [ATR on GitHub](https://github.com/Agent-Threat-Rule/agent-threat-rules): Open standard for AI agent threat detection
- 61 YAML-based rules across 9 categories: prompt injection, tool poisoning, skill compromise, context exfiltration, agent manipulation, privilege escalation, excessive autonomy, data poisoning, model security
- npm: agent-threat-rules@0.3.1

## Company

- [About](https://panguard.ai/about): Mission -- AI agent security for everyone
- [Blog](https://panguard.ai/blog): Engineering, threat intelligence, and product updates
- [GitHub](https://github.com/panguard-ai/panguard-ai): Full source code (MIT License)