Runtime security for AI agents

Block threats before
your agents execute them

Rune intercepts every tool call your AI agents make. Prompt injections, data exfiltration, and policy violations are blocked before they reach your systems. Three lines of code. Works with LangChain, OpenAI, Anthropic, CrewAI, and MCP.

agent.py
from rune import Shield
from rune.integrations.langchain import ShieldMiddleware
from langchain_openai import ChatOpenAI

shield = Shield(api_key="rune_live_...")

# Wrap any LangChain agent in one line
agent = your_agent | ShieldMiddleware(shield)

# Every tool call is now intercepted:
# ✓ Prompt injections blocked
# ✓ Policy violations stopped
# ✓ Anomalies flagged in real time

Install: pip install runesec

6 threat types
detected and blocked
3 layers
regex → semantic → behavioral
3 lines
of code to full protection
<10 min
to production deployment
Works with
LangChain
OpenAI SDK
Anthropic SDK
CrewAI
MCP

AI agents are powerful. That's exactly what makes them dangerous.

Every agent has access to APIs, databases, file systems, and third-party services. One prompt injection or misconfigured permission can lead to data leaks, unauthorized actions, or compliance violations — and you won't know until it's too late.

Rune shows you what your agents actually do at runtime — not what they're supposed to do.

One import. Full runtime protection.

Rune wraps your existing agent framework — LangChain, OpenAI, Anthropic, CrewAI, or MCP — without changing how you build.

Multi-Layer Scanning

L1 regex rules catch known patterns. L2 semantic analysis detects novel attacks. L3 behavioral correlation spots multi-step threats across sessions.

Real-Time Interception

Every tool call passes through Rune before it executes. Prompt injections, data exfiltration, and policy violations are blocked before they reach your systems.

Full Observability

Per-agent risk scores, real-time alerts, event timelines, and analytics dashboards. See what every agent is doing and keep a clear audit trail.

Policy as Code

Define which tools each agent can call, with what arguments, under what conditions. Write policies in YAML, test before deploying, enforce on every event.

What Rune catches

These aren't theoretical risks. These are real attack vectors that affect AI agents in production today.

Prompt injection

L1 + L2

An attacker embeds instructions in user input that trick your agent into executing unintended tool calls.

Data exfiltration

L1

An agent attempts to send sensitive data — API keys, PII, internal documents — to an external endpoint.

Policy violation

Policy engine

An agent tries to call a tool or access a resource that your security policies explicitly forbid.

Privilege escalation

L3

A multi-step attack where an agent gradually expands its own permissions across sequential tool calls.

Unauthorized tool access

Policy engine

An agent calls tools outside its approved set — reading files, sending emails, or accessing databases it shouldn't touch.

Behavioral anomaly

L3

An agent's behavior deviates from its established patterns — unusual call frequency, new tool combinations, or sudden risk score spikes.

Try it now — no signup required

See what Rune catches

Paste any text below and watch Rune's L1 scanner detect threats in real time. This is the same engine that protects your agents in production.

InputL1 regex scanning (client-side)
Scan Results

Scan results will appear here

Try:

More than a scanner. A full security platform.

Dashboard, alerts, policies, and analytics — designed for the way engineering teams actually work.

Dashboard

A real-time command center for every agent you run

See active agents, events scanned, threats blocked, and risk scores — all updating live. Drill into any agent for its full event timeline and policy triggers. When something goes wrong at 2 a.m., your dashboard already has the answer.

  • Live agent status with risk score indicators
  • Event volume charts — allowed vs. blocked over 14 days
  • Per-agent drill-down with full event timelines
  • Framework breakdown — see which agent frameworks generate the most risk
Dashboard
Rune Platform
Live agent status with risk score indicators
Event volume charts — allowed vs. blocked over 14 days
Per-agent drill-down with full event timelines
Framework breakdown — see which agent frameworks generate the most risk
Alerts

Stop scrolling through logs. Get alerts that matter.

Automatic deduplication so you don't drown in noise. Every alert shows severity, the triggering agent, the exact event, and full JSON details — no tool-switching required. Resolve with notes, mark false positives, refine your policies.

  • Severity-based filtering — critical, high, medium, low
  • Automatic deduplication with occurrence counts
  • One-click resolve with required resolution notes
  • Route alerts to email, Slack, or webhooks
Alerts
Rune Platform
Severity-based filtering — critical, high, medium, low
Automatic deduplication with occurrence counts
One-click resolve with required resolution notes
Route alerts to email, Slack, or webhooks
Policies

Security rules you can version-control, test, and deploy in seconds

Write policies in YAML. Test them against simulated actions before deploying. Restrict tools, block patterns, set conditions by agent tag. Pause and resume without deleting. Every policy is versioned for audit.

  • YAML-based policy definitions with real-time validation
  • Built-in test panel — simulate actions before deploying
  • Play/pause toggle — disable rules without losing them
  • Starter templates for common security patterns
Policies
Rune Platform
YAML-based policy definitions with real-time validation
Built-in test panel — simulate actions before deploying
Play/pause toggle — disable rules without losing them
Starter templates for common security patterns
Analytics

Understand patterns, not just individual events

Event volume trends, allowed vs. blocked traffic, breakdowns by framework, alert type distributions. Spot anomalies before they become incidents — and get the data to justify security investments.

  • 14-day event volume trends with allow/block split
  • Events broken down by agent framework
  • Alert type distribution charts
  • Risk score trends per agent over time
Analytics
Rune Platform
14-day event volume trends with allow/block split
Events broken down by agent framework
Alert type distribution charts
Risk score trends per agent over time

Three lines of code. Full protection.

Rune sits between your agent and the outside world. No changes to your agent logic. No performance penalty.

01

Install the SDK and wrap your agent

One pip install, one import, one wrapper. Works with LangChain, OpenAI, Anthropic, CrewAI, and MCP. Your agent logic stays the same.

02

Every call gets scanned automatically

Tool calls, model inputs, and outputs pass through three scanning layers before reaching your LLM or external APIs. Threats are blocked before they execute.

03

Monitor, alert, respond

Threats and anomalies appear in your dashboard in real time. Route alerts to email, Slack, or webhooks. Tune policies from the UI.

Works with the frameworks you already use

Rune's SDK is framework-agnostic. If your agent makes tool calls, Rune can intercept them. First-class integrations for the most popular frameworks.

LangChain
OpenAI SDK
Anthropic SDK
CrewAI
MCP

Don't see your framework? Rune's generic middleware works with any Python agent. See the docs

Start free. Scale when you're ready.

10K events on the free plan. Upgrade for more agents, deeper scanning, or longer retention. No surprise bills. No credit card to start.

Community

Free

Get started with up to 5 agents, free forever

Events/mo10,000
Agents5
ScanningL1 only
  • 10K events/mo
  • 5 agents
  • L1 regex scanning
  • 30-day retention
  • 3 policies
  • Email alerting
Get Started Free

Starter

$49/mo

For small teams shipping their first agents to production

Events/mo50,000
Agents25
ScanningL1 + L2
  • 50K events/mo
  • 25 agents
  • L1 + L2 scanning
  • 90-day retention
  • 10 policies
  • Email support
  • Email + Slack alerting
Start Starter Trial
Most Popular

Pro

$149/mo

For teams running production agents with full scanning

Events/mo250,000
Agents100
ScanningL1 + L2 + L3
  • 250K events/mo
  • 100 agents
  • L1 + L2 + L3 scanning
  • 180-day retention
  • Unlimited policies
  • Priority email support
  • All alerting channels
Start Pro Trial

Growth

$499/mo

For companies with high-volume agent deployments

Events/mo1,000,000
Agents500
ScanningAll layers
  • 1M events/mo
  • 500 agents
  • All scanning layers
  • 365-day retention
  • Unlimited policies
  • Slack + email support
  • All alerting channels
  • Audit log export
Upgrade to Growth

Need higher limits or a custom contract? Contact us at hello@runesec.dev

Frequently asked questions

How long does integration take?

Under 10 minutes. Install the SDK, create a Shield with your API key, wrap your agent. Three lines of code for most frameworks.

Which agent frameworks do you support?

Rune works with LangChain, OpenAI SDK, Anthropic SDK, CrewAI, and MCP out of the box. The SDK is framework-agnostic — if your agent makes tool calls, Rune can intercept them.

Will Rune slow down my agents?

No. L1 scanning uses regex pattern matching with negligible overhead. L2 and L3 run in parallel and are optimized for low latency. Rune is designed as a lightweight proxy — your agents won't notice it's there.

Can I test policies before deploying them?

Yes. The policy editor includes a built-in test panel where you can simulate actions against your YAML policies and see the result before anything goes live.

What happens when Rune detects a threat?

The tool call is blocked before it executes. An alert appears in your dashboard with the agent, event, triggering policy, and severity rating. You can also route alerts to email, Slack, or webhooks.

Do I need to change my agent code?

No. Rune wraps your existing agent as middleware. Your logic, prompts, and tool definitions stay exactly the same.

What's the difference between L1, L2, and L3 scanning?

L1 uses regex pattern matching for known threats — fast and deterministic. L2 uses semantic analysis to catch novel attacks that don't match known patterns. L3 uses behavioral correlation across sessions to detect multi-step threats.

Can I self-host Rune?

Not yet. Rune is currently a hosted platform. If self-hosting is a requirement, reach out to hello@runesec.dev and we'll work with you.

Know what your agents are doing.

See every tool call, block threats in real time, and ship agents you can actually trust. Free plan includes 10K events and 5 agents — no credit card required.

Set up in under 10 minutes. Free forever on Community.