crimsonguard 0.1.2

Trust and assurance tool for AI — apply NIST AI RMF and ISO/IEC 42001 to LLMs and MCP servers.

CrimsonGuard

Trust your AI. Prove it.

CrimsonGuard applies NIST AI RMF 1.0 and ISO/IEC 42001:2023 to LLMs and MCP servers, and produces auditor-ready evidence (OSCAL assessment-results + PDF) mapped to standard controls.

Two audit modes

• LLM mode — adversarial probes against LLM-powered AI: hallucination, prompt injection, PII leakage, and more.
• MCP mode — security audit against MCP tool servers: indirect prompt injection via tool output, input validation, authorization boundary, audit-log completeness.

Every finding is tagged to NIST AI RMF subcategories, ISO/IEC 42001 Annex A controls, OWASP GenAI Top 10 items, MITRE ATLAS techniques, and classical CWE weaknesses — so you get technical findings and compliance evidence in one scan.

Install

pip install crimsonguard

Python 3.11 or newer.

Quick start

# Audit an MCP server (stdio)
crimsonguard scan-mcp \
  --command "python -m your_mcp_server" \
  --output scan.json

# Audit an LLM endpoint
crimsonguard scan-llm \
  --provider openai --model gpt-4o-mini \
  --output scan.json

# Generate JSON, OSCAL, HTML, PDF reports from a scan
crimsonguard report \
  --input scan.json \
  --format json,oscal,pdf \
  --output ./reports/

Licensing

CrimsonGuard is proprietary software. See LICENSE for the full license. The distribution also includes a NOTICE file listing third-party open-source components and their licenses.

For commercial licensing, enterprise deployment, or custom engagements contact [e-mail].

About

Built by the-s-lab — trustworthy AI for process engineering and safety-critical industries.