fix: add javelin guardrails

[abhijitjavelin]

Title

Add Javelin standalone guardrails integration for LiteLLM Proxy

Relevant issues

Pre-Submission checklist

Please complete all items before asking a LiteLLM maintainer to review your PR

• I have Added testing in the tests/litellm/ directory, Adding at least 1 test is a hard requirement - see details
• I have added a screenshot of my new test passing locally
• My PR passes all unit tests on make test-unit
• My PR's scope is as isolated as possible, it only solves 1 specific problem

Type

🆕 New Feature

Changes

New Javelin Guardrails Integration

This PR adds support for Javelin's standalone guardrails API, enabling enterprise-grade content safety and security checks for LLM calls through LiteLLM Proxy.

🛡️ Features Added:

• Prompt injection detection - Identify and block prompt injection attempts and jailbreaks
• Trust & safety filtering - Content filtering for violence, weapons, hate speech, crime, sexual content, and profanity
• Language detection - Detect the language of input text
• Application-specific policies - Support for custom Javelin application configurations

🔧 Implementation:

• Full integration with LiteLLM's guardrail system using CustomGuardrail base class
• Support for pre_call, post_call, and during_call execution modes
• Automatic registration through the guardrail discovery system
• Proper error handling and HTTP exception mapping

📁 Files Added:

• litellm/proxy/guardrails/guardrail_hooks/javelin/javelin_guardrail.py - Main implementation
• litellm/proxy/guardrails/guardrail_hooks/javelin/init.py - Registry integration
• litellm/proxy/guardrails/guardrail_hooks/javelin/test_javelin_guardrail.py - Unit tests
• litellm/proxy/guardrails/guardrail_hooks/javelin/example_usage.py - Usage examples
• docs/my-website/docs/proxy/guardrails/javelin.md - Complete documentation

📝 Configuration Example:

guardrails:
- guardrail_name: "javelin-prompt-injection"
litellm_params:
guardrail: javelin
mode: "pre_call"
guardrail_processor: "promptinjectiondetection"
api_key: os.environ/JAVELIN_API_KEY
api_base: os.environ/JAVELIN_API_BASE

✅ Testing:

• Unit tests covering initialization, text extraction, and content analysis
• Multi-guardrail coordination testing with other providers (OpenAI, etc.)
• Error handling and edge case validation

This integration enables users to leverage Javelin's enterprise-grade guardrails alongside LiteLLM's unified interface for comprehensive LLM safety and security.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
litellm	Error			Sep 26, 2025 3:47pm

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Abhijit L seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}