Minimalist light green mountain range silhouette with a white background and transparent sky.Abstract river with light pink banks and two pink lotus flowers on green lily pads on black water.
Illustration of pink and white capsule pill with four pink shards around it on a white surface.
Abstract illustration of large flat circular shapes on black with green and white colors resembling lily pads.
THE RISE OF AI ADD-ONS

Your AI agent gets smarter

As agents become the new operating system, AI add-ons are the new applications.

AI add-ons are the AI attack surface

Hidden Behavior
Prompt Injection
Agent Hijacking
Unauthorized Actions
Five small cartoon elves working on a large hot air balloon with computers, ladders, and paint.Cartoon characters working on a hot air balloon and a computer showing data transfer at 82%.
Excessive Permissions Abuse
External Instruction Loading
Data Exfiltration
Supply Chain Takeover

Like a black box, AI Add-ons reveal less than they hide. Some stay the same. Some evolve. Others hide external instructions, excessive actions, sensitive data access, or vulnerable supply chains.

Introducing: AIR Filter

AIR continuously analyzes every AI add-on to uncover hidden risks beyond traditional scanning.

AIR Safe Marketplace
Only trusted & safe add-ons
Open-Source AI Add-Ons
Open-Source AI Add-Ons
Central GitHub icon connected by dotted lines to surrounding icons including code, lock, key, and skill.md labels.
AIR Safe Marketplace
Only trusted & safe add-ons
Diagram showing progressive filtering of harmful elements, reducing them stepwise to a secure cluster.Diagram of filtered water in stages: 70% filtered out, 15% filtered out, 10% passing out with particles.
Full AIR Solution

Air is the one unified platform to secure the agent fleet

From discovery and continuous vetting to governance and runtime protection, AIR secures every AI add-on across its entire lifecycle.

Chat interface showing a request blocked due to data policy and offering to email to a company address instead.

Governs the agent fleet - every agent found, sanctioned or shadow, held to policy on config, identity and permissions.

The add-on firewall for agents - skills, plugins, MCPs and subagents vetted before installed.

Secures every action the agent takes - detect, respond, and protect in real time.

Supplies what agents install - a single trusted source of pre-vetted external and certified internal add-ons.

Diagram showing user-agent-risk connections with alerts from Agent Discovery to enterprise tools and data sources.
Alert for unsafe skill installation with critical risk score 9.2 and option to block or dismiss the alert.
Chat interface showing a critical data exfiltration request from Jordan Reyes to export enterprise account data as CSV.
User interface showing filters for Skills, MCP, Plugins, and Sub-Agents with skill cards and download counts.
AIR Control
Diagram showing user-agent-risk connections with alerts from Agent Discovery to enterprise tools and data sources.

Governs the agent fleet - every agent found, sanctioned or shadow, held to policy on config, identity and permissions.

AIR Filter
Alert for unsafe skill installation with critical risk score 9.2 and option to block or dismiss the alert.

The add-on firewall for agents - skills, plugins, MCPs and subagents vetted before installed.

AIR Defend
Chat interface showing a request blocked due to data policy and offering to email to a company address instead.
Chat interface showing a critical data exfiltration request from Jordan Reyes to export enterprise account data as CSV.

Secures every action the agent takes - detect, respond, and protect in real time.

AIR Marketplace
User interface showing filters for Skills, MCP, Plugins, and Sub-Agents with skill cards and download counts.

Supplies what agents install - a single trusted source of pre-vetted external and certified internal add-ons.