Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Latest commit

 

History

History
94 lines (70 loc) · 3.45 KB

File metadata and controls

94 lines (70 loc) · 3.45 KB

Prisma AIRS Integrations

Documentation repository for integrating Prisma AIRS (AI Runtime Security) with third-party platforms.

Repository Purpose

This is a documentation-first repository - primarily markdown guides, configuration examples, and sample code for integrating Prisma AIRS with AI gateways, LLM platforms, and automation tools.

Structure

├── Anthropic/          # Anthropic integrations (Claude Code hooks, MCP, skill)
├── Microsoft/          # Microsoft integrations (Azure APIM)
├── Google/             # Google Cloud integrations (Apigee)
├── Kong/               # Kong custom plugin & request callout
├── LiteLLM/            # LiteLLM proxy integration
├── n8n/                # Workflow automation
├── Portkey/            # AI Gateway & observability
└── TrueFoundry/        # AI Gateway

Each integration follows the template in CONTRIBUTING.md:

  • README.md - Main setup guide with Coverage table
  • config/ or examples/ - Sample configurations
  • scripts/ - Setup/test scripts

Coverage Table Format

Every integration README must include a Coverage section after the title/description. Use this format:

## Coverage

> For detection categories and use cases, see the [Prisma AIRS documentation](https://pan.dev/prisma-airs/api/airuntimesecurity/usecases/).

| Scanning Phase | Supported | Description |
|----------------|:---------:|-------------|
| Prompt || Scans user prompts before sending to LLM |
| Response || Scans LLM responses before returning to user |
| Streaming || Real-time scanning of streamed responses |
| Pre-tool call || Scans before tool/function execution |
| Post-tool call || Scans tool/function results |

Support icons:

  • ✅ Full support
  • ⚠️ Partial/conditional support (describe conditions in Description column)
  • ❌ No support

Column definitions:

  • Prompt: Scans user input before sending to LLM
  • Response: Scans LLM output before returning to user
  • Streaming: Real-time scanning of streamed LLM responses
  • Pre-tool: Scans before tool/function calls are executed
  • Post-tool: Scans results returned from tool/function calls

Working in This Repo

Adding New Integrations

  1. Create directory: PlatformName/
  2. Follow structure in CONTRIBUTING.md (line 74-86)
  3. Include README with: Prerequisites, Configuration Steps, Validation
  4. Update root README.md table with new integration

Technical Requirements for All Integrations

All integrations MUST:

  1. Set app_name field in AIRS requests (format: <VENDOR>-<CUSTOMER_APP>)
  2. Use tr_id to pass unique identifiers from the platform when available
  3. Never include real credentials - use placeholders like YOUR_API_KEY_HERE

Commit Convention

docs: add [Platform] integration guide
fix: correct [Platform] configuration syntax
feat: add automated setup script for [Platform]
test: add validation examples for [Platform]

Key Files

File Purpose
README.md Integration index and overview
CONTRIBUTING.md Contribution guidelines and templates

External Resources