Add ClawGuard — MCP prompt injection scanner (132 patterns, 14 languages)

[joergmichno]

ClawGuard is a defensive MCP security scanner:

• 132 regex patterns, 14 languages, F1=97.4%
• Detects prompt injection, tool poisoning, data exfiltration
• EU AI Act compliance mapping (unique)
• Sub-10ms, MIT licensed

GitHub: https://github.com/joergmichno/clawguard-mcp
Blog: https://prompttools.co/blog/6-real-mcp-security-breaches

[joergmichno]

Update: ClawGuard has grown significantly since this submission — now at 207 patterns across 15 languages with F1=98.0% and 264 test cases. We've also added CoSAI/OASIS MCP Security Taxonomy mapping (8/12 categories covered). Happy to update the description if you'd like before merging.

[joergmichno]

Update (22.03.2026): ClawGuard has grown significantly since the initial submission:

• 200 patterns (was 132) across 15 languages
• F1=97.3% on 265 adversarial test cases
• 24 new red-team patterns (gaslighting, roleplay evasion, reversed text, hypothetical framing, etc.)
• EU AI Act compliance scanning + full OWASP LLM/Agentic/MCP Top 10 mapping
• Free API: https://prompttools.co/shield
• Scanned 11,529 MCP servers from public registries

GitHub: https://github.com/joergmichno/clawguard (MIT)

[esxr]

Another one worth adding to the list: operant-mcp. Open source MCP server with 51 security testing tools. Covers SQLi, XSS, command injection, SSRF, path traversal, PCAP forensics, memory forensics, recon, malware analysis, CloudTrail, and more. MIT license. https://github.com/operantlabs/operant-mcp

[joergmichno]

Final update (26.03): Now at 245 patterns, 15 languages, F1=99.0%. Also added: GitHub Action (joergmichno/clawguard-action@v1), MCP Server, and EU AI Act compliance reporting. Full details on the repo.

[vincenzodentamaro]

Your integration of EU AI Act compliance reporting and expanded security patterns with ClawGuard highlights the critical need for robust AI safety and regulatory alignment. Geodesia G-1 provides an enterprise-grade platform that not only ensures such compliance but also offers comprehensive AI safety governance, bridging the gap between advanced security measures and regulated deployments.

Geodesia G-1 is the enterprise-ready solution for this regulatory gap. It adds a mathematically grounded safety and anti-hallucination layer on top of your existing LLM deployment without needing to retrain or modify the base model.

Key features for compliance and deployment:

• Runtime Audit Logs & Watermarking
• FRIA (Fundamental Rights Impact Assessment) support
• Human-in-the-loop oversight workflows
• Automated Compliance Report Generation

Performance Metrics (Published):

• AUROC > 0.90 for safety detection
• AUROC > 0.96 for hallucination detection
• <35 ms total added overhead
• Coverage for 13+ regulatory frameworks (including EU AI Act)

Geodesia provides the necessary observability and safety layer to make AI models deployable in highly regulated environments with measurable trust.

Website: https://www.geodesia.ai