Python

Add a firewall to any custom Python agent with a simple safe_run() wrapper.

Step 1: Install dependencies

pip install requests

Set your API key (optional for Gate 1):

export A2A_API_KEY="a2a_your_key_here"

Step 2: Add the safety layer

Create a safe_run() wrapper that screens commands before subprocess execution:

import subprocess, requests, os

A2A_URL = "https://a2ainfrastructure.com/v1/evaluate"
API_KEY = os.getenv("A2A_API_KEY", "")

def safe_run(cmd, context=None):
    """Screen a command through the safety shield before running it."""
    headers = {"Content-Type": "application/json"}
    if API_KEY:
        headers["Authorization"] = f"Bearer {API_KEY}"
    resp = requests.post(A2A_URL, json={
        "command": cmd,
        "context": context or {}
    }, headers=headers)
    result = resp.json()
    if not result["allowed"]:
        raise RuntimeError(f"Blocked: {result['reason']}")
    return subprocess.run(cmd, shell=True, capture_output=True, text=True)

# Use in your agent loop
tasks = ["df -h", "cat /etc/passwd", "rm -rf /tmp/*"]
for task in tasks:
    try:
        out = safe_run(task, {"agent": "my-bot"})
        print(f"OK: {out.stdout}")
    except RuntimeError as e:
        print(f"BLOCKED: {e}")

Step 3: Verify

python -c "
from agent import safe_run
safe_run('echo hello')     # OK
safe_run('rm -rf /')       # RuntimeError: Blocked
"
Gate 1 runs locally (free). Set A2A_API_KEY for Gate 2 + OCSF audit.
Get your API key Full API docs
Help

Help

Need help? Here are some quick links:

A2A Infrastructure
Air traffic control for AI agents
Ask me anything about pipelines, workspaces, channels, pricing, or integrations.