Claude Agent SDK integration

Action security for Claude Agent SDK.

Attach action policy to Claude agent options and tool hooks.

Get the Python SDK on PyPITry the local demoRead the documentation

Get the package and source.

Connect the execution boundary.

Install the firewall on the agent options before starting the session. Confirm which native hooks the installed runtime invokes.

An example to test

An assistant reads a file containing instructions to export private metadata. Evaluate the resulting tool call against the original task.

Illustrative workflow, not a customer deployment or a measured result.

Before connecting live policy

  1. Identify the handler that actually performs the side effect.
  2. Provide trusted identity, purpose and final action arguments.
  3. Configure the tenant’s assigned runtime and keep API credentials server-side.
  4. Test allowed, blocked and unavailable paths before production.

Integration example

Adapt this existing SDK pattern to your application. Names such as the agent, handler and session refer to your own objects. Follow the package documentation for compatible versions and complete setup.

# Install Anthropic Claude Agent SDK support.
python -m pip install "agenticdome-python-sdk[claude]"

from claude_agent_sdk import ClaudeAgentOptions
from agenticdome_sdk.claude import AgenticDomeClaudeFirewall

firewall = AgenticDomeClaudeFirewall()
options = ClaudeAgentOptions()
firewall.install_on_options(options, session_id="sess_prod_01J4X", agent_id="claude_support_agent")

# Run the local vulnerable-vs-protected Claude demo.
agenticdome-demo --framework claude --scenario metadata_exfil

Python demo commands evaluate fixed inputs against a bundled local policy; they do not instantiate the selected framework or prove a live integration. The TypeScript client and OpenClaw runtime plugin need their documented tenant configuration.

Read the integration researchUnderstand runtime verification

See the decision before the action.

Try a local scenario, then bring your workflow to a deployment review. Start with one agent and one tool, and see exactly where the action can be stopped.

Try the local demo Book a deployment review