Reference

Security Glossary

Essential security concepts explained. From Prompt DLP to secret sprawl, understand the terminology behind modern AI-native security.

Prompt DLP

Data Loss Prevention for AI prompts — intercept, scan, and block secrets or PII before they reach LLM providers.

Byte-Pair Encoding applied to secret detection — tokenize text into subwords and identify high-entropy subsequences that indicate leaked credentials.

Machine credentials such as service accounts, API keys, OAuth tokens, and managed identities that outnumber human users 100 to 1.

The uncontrolled proliferation of credentials across code repositories, logs, collaboration tools, and CI/CD pipelines.