Jailbreak

A prompt or interaction pattern designed to bypass model restrictions. Success against one version may fail against another, so jailbreak rates require dated and versioned tests.

Security and privacy

Where an agent, a tool or a page can be turned against its operator, and the controls that limit the damage without proving safety.

Indirect prompt injection · Direct prompt injection · Data exfiltration · Sensitive information disclosure · Secret leakage · Insecure output handling · Excessive agency · Least privilege · Trust boundary · Input validation · Output validation · Allowlist · Denylist · SSRF · RCE · Credential scope · OAuth scope · API key · Bearer token · mTLS · Rate-limit policy · Audit log · PII · Data minimization · Purpose limitation · Retention period · Data residency · Tenant isolation · Threat model

← Direct prompt injection · Data exfiltration →

See it in the full glossary · 579 terms across 19 areas. Scan a site to see which of these apply to it.