AgentGuard Python SDK
Python SDK for integrating with AgentGuard - AI Agent governance and monitoring platform.
Features
- ✅ Zero Code Changes - Drop-in replacement for OpenAI client
- ✅ Unified Client - One client for all features (LLM + Approvals)
- ✅ Transparent Proxy - All requests routed through AgentGuard
- ✅ Integrated Approval Management - Built-in approval status query and submission
- ✅ Business API Interception - Monitor all API calls
- ✅ Cost Tracking - Automatic cost monitoring
- ✅ Policy Enforcement - Apply governance policies
- ✅ Approval Workflows - Handle high-risk operations
Installation
pip install agentguard-zhx
Quick Start
Basic Usage
from agentguard import AgentGuardOpenAI
# Initialize client
client = AgentGuardOpenAI(
agentguard_url="http://localhost:8080",
agent_api_key="ag_xxx"
)
# Use standard OpenAI API
response = client.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)
Approval Status Query
Query approval status using integrated approval management:
from agentguard import AgentGuardOpenAI
# One client for everything
client = AgentGuardOpenAI(
agentguard_url="http://localhost:8080",
agent_api_key="ag_xxx"
)
# LLM calls
response = client.chat.completions.create(...)
# Approval management (integrated)
status = client.approvals.get_status("approval_id")
if status.is_approved:
print("Approved:", status.execution_result)
elif status.is_rejected:
print("Rejected:", status.remark)
# Submit approval reason
client.approvals.submit_reason("approval_id", "Need to delete test data")
Environment Variables
# Set environment variables
export AGENTGUARD_URL="http://localhost:8080"
export AGENTGUARD_API_KEY="ag_xxx"
# Use without explicit configuration
from agentguard import AgentGuardOpenAI
client = AgentGuardOpenAI() # Loads from environment
Business API Interception
from agentguard import enable_agentguard
import requests
# Enable global interception
enable_agentguard(
agentguard_url="http://localhost:8080",
agent_api_key="ag_xxx",
intercept_patterns=[r"https://api\.example\.com/.*"]
)
# All matching requests go through AgentGuard
response = requests.get("https://api.example.com/data")
Configuration
AgentGuardConfig
from agentguard import AgentGuardConfig, AgentGuardOpenAI
config = AgentGuardConfig(
agentguard_url="http://localhost:8080",
agent_api_key="ag_xxx"
)
client = AgentGuardOpenAI(config=config)
Examples
See the examples/ directory for more examples:
basic_usage.py- Basic OpenAI integrationapproval_polling.py- Query approval statusbusiness_api.py- Business API interception
API Reference
AgentGuardOpenAI
Drop-in replacement for OpenAI client with integrated approval management.
Parameters:
agentguard_url(str): AgentGuard server URLagent_api_key(str): AgentGuard API keyconfig(AgentGuardConfig, optional): Configuration object
Integrated Approval Management:
Access via client.approvals:
get_status(approval_id)- Query approval status by IDsubmit_reason(approval_id, reason)- Submit approval reason/justification
Returns: ApprovalStatusResponse with:
status- ApprovalStatus enum (PENDING/APPROVED/REJECTED/EXPIRED)execution_result- Result if approved and executedremark- Rejection reason if rejectedis_pending,is_approved,is_rejected,is_expired- Helper properties
enable_agentguard()
Enable global request interception for the requests library.
Parameters:
agentguard_url(str): AgentGuard server URLagent_api_key(str): AgentGuard API keyintercept_patterns(List[str], optional): URL patterns to intercept (regex)
Development
# Install development dependencies
pip install -e ".[dev]"
# Run tests
pytest
# Format code
black agentguard/
# Lint code
ruff check agentguard/
License
MIT License
Support
Metadata
Release files for agentguard-zhx 0.0.3
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agentguard_zhx-0.0.3.tar.gz | 14.5 kB | Without provenance | Details for agentguard_zhx-0.0.3.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agentguard_zhx-0.0.3-py3-none-any.whl | Python 3 | none | any | Details for agentguard_zhx-0.0.3-py3-none-any.whl |
Total release size: 31.3 kB
Release files / agentguard_zhx-0.0.3.tar.gz
| Download URL | agentguard_zhx-0.0.3.tar.gz |
|---|---|
| Size | 14.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
1456a6071e7b45d7c7767e57a37f7c120e30e36091569286020c61b0351643c9
|
|
BLAKE2b-256 checksum How to use checksums |
7a9f23995a9f4ba59bf38079a43f8a27f7c18e8e3c0c3b3f92a202b06aeb20ed
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.6
|
Release files / agentguard_zhx-0.0.3-py3-none-any.whl
| Download URL | agentguard_zhx-0.0.3-py3-none-any.whl |
|---|---|
| Size | 16.9 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
0c9f1e8bd1c874ec974de45c99c46d0d5b67f57ee332f7b14028c0e4cb0d855e
|
|
BLAKE2b-256 checksum How to use checksums |
012bbfa51d523399c1635cf8cd6fa570a2a089d98cc4cc2044130606a8608710
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.6
|
Security for agentguard-zhx 0.0.3
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.