AI Agent PQC Risk Audit
Scan LangChain, LlamaIndex, and AI API call patterns for quantum-unsafe crypto, hardcoded secrets, and missing evidence.
Please login to run AI Agent PQC scans and view your audit history.
• Parses agent/framework code (LangChain, LlamaIndex, Autogen)
• Detects TLS < 1.3 on AI API calls (OpenAI, Anthropic, Gemini, Azure OpenAI, xAI/Grok)
• Flags hardcoded API keys
• Checks for crypto-agility (can we swap to PQC/hybrid)
• Ensures calls are logged (CISA Secure AI guidance)
Output goes to: Dashboard → PDF → JSON → CSV (your standard output trio).
| Provider \ Framework | LangChain | LlamaIndex | Autogen | Custom | Direct API |
|---|---|---|---|---|---|
| OpenAI | |||||
| Anthropic | |||||
| Self-hosted LLaMA |
T = Transport PQ-ready (TLS1.3+) • K = Key/secret handling ok • C = Crypto-agile • L = Logged/evidence
Send agent/framework usage and we'll score it:
POST /api/audits/ai-agent-pqc
{
"frameworks": ["langchain"],
"code_snippets": ["import langchain", "OPENAI_API_KEY='sk-...'"],
"calls": [
{
"provider": "openai",
"url": "https://api.openai.com/v1/chat/completions",
"tls_version": "TLS1.2",
"auth": "hardcoded",
"logged": false,
"crypto_agile": false
}
]
}
Response format:
{
"module": "ai_agent_pqc",
"score": 45,
"findings": [
{
"id": "TLS_BELOW_1_3",
"severity": "high",
"text": "Connection uses TLS version lower than 1.3...",
"weight": 20
}
],
"matrix": { ... },
"report_id": "ai-agent-pqc-abc123"
}
Documentation for this module is available throughout the platform:
• Home → overview and getting started
• Product → full module catalog
• AI Defense Cloud → parent module hub
• FAQ → methodology and scoring