AuditaAI
Back to vendors

AI security vendor

Anthropic

Commercial parent organization

Official website

2

Offerings

3

Reviewed claims

4

Capabilities

Domains

Agent SecurityRuntime Protection

Actions

PREVENT

Offerings

Open an offering for reviewed behavior and evidence

Anthropic · module

Constitutional Classifiers

1 claims+

Model safeguard classifiers designed to defend Claude against jailbreak attempts.

Official source
Prompt and Response Guardrails
PREVENT

Anthropic applies constitutional classifier safeguards to detect and block jailbreak attempts against Claude model behavior.

Source

Anthropic · module

Claude Code Security Architecture

2 claims+

Permission and sandbox controls that constrain Claude Code actions on local systems.

Official source
Agent-Human OversightAgent Tool Governance
PREVENT

Claude Code requires permission for state-changing file, command, and network actions before those actions execute.

Source
Agent Execution Isolation
PREVENT

Claude Code constrains command execution with operating-system sandbox boundaries that limit filesystem and network access.

Source