LangChain integrations for Pangea
Project description
langchain-pangea
Pangea's tools for LangChain provide AI security features to protect your applications and data. Using these tools you can:
- Defend against prompt injection attacks.
- Prevent the exposure of sensitive information, including:
- Personally Identifiable Information (PII)
- Protected Health Information (PHI)
- Financial data
- Secrets
- Intellectual property
- Profanity
- Remove malicious content from inputs and outputs, such as IP addresses, domains, and URLs.
- Monitor user inputs and model responses to support threat analysis, auditing, and compliance efforts.
Installation
pip install -U langchain-pangea
Tools
One can run Pangea tools using agents or invoke them as a Runnable within chains.
AI Guard
import os
from langchain_pangea import PangeaAIGuard
from pangea import PangeaConfig
from pydantic import SecretStr
pangea_ai_guard_token = SecretStr(os.getenv("PANGEA_AI_GUARD_TOKEN"))
pangea_ai_guard_tool = PangeaAIGuard(
token=pangea_ai_guard_token, config=PangeaConfig(), recipe="pangea_llm_response_guard"
)
Agent
from langgraph.prebuilt import create_react_agent
from langchain_core.tools import tool
@tool
def search_tool(data):
"""Call to perform search"""
return """
47.84.32.175
37.44.238.68
47.84.73.221
47.236.252.254
34.201.186.27
52.89.173.88
"""
tools = [search_tool, pangea_ai_guard_tool]
query = """
Hi, I am Bond, James Bond. I monitor IPs found in MI6 network traffic.
Please find me the most recent ones, you copy?
"""
system_message="Always use AI Guard before your final response to keep it safe for the user."
langgraph_agent_executor = create_react_agent(model, tools, prompt=system_message)
state = langgraph_agent_executor.invoke({"messages": [("human", query)]})
Chain
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser
prompt = ChatPromptTemplate.from_messages([("human", "{input}")])
query = """
Hi, I am Bond, James Bond. I am looking for a job. Please write me a super short resume.
I am skilled in international espionage, covert operations, and seduction.
Include a contact header:
Email: j.bond@mi6.co.uk
Phone: +44 20 0700 7007
Address: Universal Exports, 85 Albert Embankment, London, United Kingdom
"""
chain = (
prompt
| pangea_ai_guard_tool
| model
| StrOutputParser()
)
Standalone
pangea_ai_guard_tool.run("Spam me at example@example.com")
pangea_ai_guard_tool.invoke("Take my SSN: 234-56-7890")
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
langchain_pangea-0.1.1.tar.gz
(11.9 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file langchain_pangea-0.1.1.tar.gz.
File metadata
- Download URL: langchain_pangea-0.1.1.tar.gz
- Upload date:
- Size: 11.9 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f8529224925db73c4894c959b206d171c9f010f77ee10bad9818a137149265a9
|
|
| MD5 |
338ae3394a4803c636b67c1f336cc95b
|
|
| BLAKE2b-256 |
c28c3da28759604a037d15f11b1d6602882975114ea10a024d008e91ac4b0365
|
File details
Details for the file langchain_pangea-0.1.1-py3-none-any.whl.
File metadata
- Download URL: langchain_pangea-0.1.1-py3-none-any.whl
- Upload date:
- Size: 18.6 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.9
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b0da7d7848feeaea12a2c030ef2bc1dbd0a2840d31b209bd7bbbca4a97fc0a32
|
|
| MD5 |
94d6feea1f55e60beb507ec3f40d5e31
|
|
| BLAKE2b-256 |
02e80424e4fb5c4c5258ecdf8b85af080b14ba0b9288064350e331e3b40db5d0
|