Set up ThreatVec

5 minutes. Shadow AI data starts flowing in step 2.

Journey

Extension

SDK

Enrich

Ready

👋

Welcome to ThreatVec

Posture scoring for your AI agent attack surface — works standalone or alongside your enforcement tools.

What are you setting up?

🧩

No code required

Install the Chrome extension

This is the fastest path to real data. Install it now and shadow AI usage starts appearing in your dashboard within minutes — no agent code needed.

✓ Detects AI sites visited by your team (ChatGPT, Gemini, Claude.ai, 80+ more)

✓ Sends PII type signals when sensitive data is typed into AI tools — never the actual values

✓ Shadow score starts populating immediately

Installation options:

⬇️

Download extension (.zip)

Load in Chrome → Extensions → Developer mode → Load unpacked

🏪

Chrome Web Store

Coming soon — pending Google review

Extension installed — shadow AI monitoring is active.

📦

2 lines of Python

Register your agent

Install + add to your entry point

Set these environment variables

Name examples: research-agent · support-bot · data-pipeline — defaults to hostname if not set

Your agent will appear in the dashboard on next restart.

🔌

Connect data sources (optional)

All optional — ThreatVec is scoring already. Each source adds a new class of signals to the posture score.

Identity & Endpoint signals

Okta Domain

API Token

Set OKTA_DOMAIN and OKTA_API_TOKEN in your .env and restart.

Client ID

Client Secret

Set CROWDSTRIKE_CLIENT_ID / CROWDSTRIKE_CLIENT_SECRET in your .env.

Pulls AI-service DNS/proxy events from your existing SIEM — no new agents needed.

Splunk URL

Splunk Token

Or configure Microsoft Sentinel: set SENTINEL_WORKSPACE_ID + SENTINEL_TENANT_ID in .env.

Enforcement & red team

Snapper URL

Read-only API Key

Generate in Snapper → Admin → API Keys

Configure URLs now — data flows in automatically when Sprint 2 ships.

Acuvity URL

Runlayer MCP Proxy URL

⏳

Computing your posture score…

From the dashboard, try

📊 Check the Posture Score — the headline number tells you how well your AI attack surface is covered

🌑 Open Shadow AI Monitor — see what AI tools your team is using outside approved channels

⚖️ Visit Settings → Score Weights to tune how Agent, MCP, and Shadow AI factors are balanced

🎯 Connect Snapper (optional) to pull red team results — the Framework Coverage section shows MITRE/OWASP mapping

SDK snippet for your team

Email to your team

View dashboard →