> ## Documentation Index
> Fetch the complete documentation index at: https://docs.definable.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Security Examples

> Examples for tool policies, rate limiting, prompt injection detection, and security audits.

## Basic Security

Enable default security configuration:

```python theme={null}
from definable.agent import Agent

agent = Agent(model="gpt-4o", security=True)
```

## Tool Policy (Allowlist)

Only allow specific tools:

```python theme={null}
from definable.agent import Agent
from definable.agent.security import SecurityConfig, ToolPolicy

agent = Agent(
    model="gpt-4o",
    tools=[search_web, delete_file, read_file],
    security=SecurityConfig(
        tool_policy=ToolPolicy(
            mode="allowlist",
            allowed_tools={"search_web", "read_file"},  # delete_file is blocked
        ),
    ),
)
```

## Tool Policy (Blocklist)

Block specific tools:

```python theme={null}
agent = Agent(
    model="gpt-4o",
    tools=[search_web, delete_file, read_file],
    security=SecurityConfig(
        tool_policy=ToolPolicy(
            mode="blocklist",
            blocked_tools={"delete_file"},
        ),
    ),
)
```

## Security Audit

Run a security audit to get a score and findings:

```python theme={null}
import asyncio
from definable.agent import Agent
from definable.agent.security import SecurityConfig, ToolPolicy

agent = Agent(
    model="gpt-4o",
    tools=[search_web],
    security=SecurityConfig(
        tool_policy=ToolPolicy(mode="allowlist", allowed_tools={"search_web"}),
    ),
)

async def main():
    report = await agent.security_audit()
    print(f"Score: {report.score}/100")
    for finding in report.findings:
        print(f"  [{finding.severity}] {finding.message}")

asyncio.run(main())
```

## Combined with Guardrails

```python theme={null}
from definable.agent import Agent
from definable.agent.security import SecurityConfig, ToolPolicy
from definable.agent.guardrail import Guardrails, max_tokens, pii_filter

agent = Agent(
    model="gpt-4o",
    security=SecurityConfig(
        tool_policy=ToolPolicy(mode="allowlist", allowed_tools={"search_web"}),
    ),
    guardrails=Guardrails(
        input=[max_tokens(500)],
        output=[pii_filter()],
    ),
)
```

<Note>
  When `security=SecurityConfig(tool_policy=...)` is set, a `ToolPolicyGuardrail` is auto-injected into the agent's guardrails. Do not duplicate it manually.
</Note>