GitHub Repo Scanning

Scan your GitHub organization or personal repos to find repositories that contain AI agent code.

What It Detects

The GitHub scanner uses the Code Search API to find repos containing AI framework imports:

PatternFrameworkLanguage
from langchainLangChainPython
from crewaiCrewAIPython
import autogenAutoGenPython
from llama_indexLlamaIndexPython
import openaiOpenAI SDKPython
import anthropicAnthropic SDKPython
from "openai"OpenAI SDKTypeScript
import dev.langchain4jLangChain4jJava

Setup

1

Create a GitHub personal access token

Go to github.com/settings/tokens and generate a classic token with the repo scope.

2

Add the integration

Go to Agent Inventory → Integrations → Add Integration, select GitHub Repos, and paste your token.

3

Set your org or username

In the Configuration section, enter your GitHub organization name or personal username. The scanner will search all repos under that owner.

4

Click Add & Scan

The scan runs immediately. Each repo with AI code patterns appears in your inventory as a discovered agent.

Indexing delay

GitHub's Code Search index can take up to 30 minutes to index newly created or recently pushed repos. If a repo doesn't appear, wait and re-scan.

What You Get

For each repo with AI code, the inventory shows:

  • Agent name — the repository name
  • Framework — the primary detected framework (LangChain, CrewAI, etc.)
  • Language — Python, TypeScript, or Java
  • Dependencies — all detected framework/SDK imports as API dependencies
  • Metadata — repository full name and detected patterns

Next Steps

After discovering an agent via GitHub scan, click Onboard on the agent's detail page to get SDK setup instructions for the detected language. Once the SDK is added, the agent transitions to full monitoring with traces, token usage, anomaly detection, and compliance tracking.