AIR raises $50M to help companies vet the skills and add-ons AI agents use
AI security startup AIR has emerged from stealth with $50 million raised across two seed rounds to help companies vet the growing array of skills, plugins and MCP servers that AI agents rely on. As businesses increasingly grant agents access to sensitive systems, a new software supply chain of add-ons and tools has formed with little of the oversight given to traditional software, leaving firms exposed to attacks in which malicious actors poison the content an agent consumes rather than attacking it directly.
Founded by Yair Saban and Niv Hoffman, both veterans of Israel's Unit 8200 intelligence corps, AIR discovers agents running within a company, continuously checks the tools and skills they use against a maintained whitelist, and blocks anything that fails its security criteria, while also flagging unapproved AI tool use by employees. The funding comprised a $10 million round led by Sequoia and a $40 million round led by Greenoaks, with angel investors including Cognition's Zach Frankel and Wiz co-founder Yinon Costica also taking part. Saban says the platform currently filters out around 27% of add-ons and skills it scans, and that AIR has more than 20 customers, roughly a quarter of them large enterprises, with strongest demand from regulated sectors such as financial services and pharmaceuticals. Competitors including Noma Security, Zenity and Astrix Security are building similar tools.
- AIR raises $50M to secure AI agents' skills, plugins and MCP tools
- Founded by ex-Israeli intelligence officers; Sequoia and Greenoaks led rounds
- Platform filters out ~27% of scanned add-ons; over 20 customers so far
New here? Start with this
AI agents are software systems that can carry out tasks with limited human input, such as searching company files or using business applications. They often gain new abilities through add-ons such as skills, plugins and MCP servers, which connect them to tools and information. These add-ons can become a route into sensitive company systems.
Companies have long checked traditional software for security risks, but many agent add-ons are newer and receive less scrutiny. A malicious or compromised add-on could expose data or influence what an agent does by feeding it harmful content. This has led to a growing focus on checking the tools agents use and limiting access to those that meet security requirements.
AIR is one of several companies developing software to find and assess these tools inside businesses. Its founders are Yair Saban and Niv Hoffman, and the company says it serves more than 20 customers, including firms in regulated sectors such as finance and pharmaceuticals. Its work matters as businesses decide how to use AI agents while managing the risks of giving them access to important systems.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Companies are giving AI agents access to sensitive systems, while the skills and add-ons those agents use can introduce risks that traditional software checks may miss. AIR’s continuous vetting and blocking could help organisations manage that emerging supply chain, especially in regulated industries where preventing a breach is a high priority.
The case against
A continuously maintained whitelist may give companies a useful control, but it could also block legitimate tools or create a false sense of security if the checks miss harmful behaviour or fail to keep pace with changes. With several firms developing similar products and AIR’s effectiveness not independently established in the report, businesses could reasonably question whether this approach justifies its cost and operational burden.