Capsule Security Exits Stealth with $7m Seed Round

Why It Matters

As AI agents become privileged, enterprise security gaps at runtime grow, making Capsule’s controls critical for safe deployment.

Key Takeaways

• $7M seed round led by Lama Partners, Forgepoint Capital
• Platform secures AI agents during execution, preventing prompt injection
• Disclosed critical bugs in Microsoft Copilot Studio, Salesforce Agentforce
• Open‑source tools ClawGuard and OpenClaw enforcer released
• Backed by former CISA director and top industry CISOs

Pulse Analysis

The enterprise AI landscape is shifting as autonomous agents move from experimental labs into production. Microsoft reports that more than 80 % of Fortune 500 firms now run active AI agents built with low‑code or no‑code platforms, turning these bots into a new class of privileged user that can read, write, and execute code at machine speed. Traditional security tools, designed for static applications, lack visibility into the fleeting moment between a prompt and an action, creating a “runtime gap” that attackers can exploit.

Capsule Security aims to close that gap with a runtime trust layer that intercepts every decision point an agent makes. The platform enforces policy checks before tool calls, preventing prompt‑injection attacks, unauthorized data exfiltration, and unintended behavior. In its stealth exit, Capsule disclosed two zero‑day vulnerabilities—ShareLeak in Microsoft Copilot Studio and PipeLeak in Salesforce Agentforce—both patched after public disclosure. The company also contributed the open‑source ClawGuard checkpoint and an OpenClaw enforcer, giving developers immediate safeguards while the seed round, led by Lama Partners and Forgepoint Capital, supplies $7 million for scaling.

With advisors such as former CISA director Chris Krebs and former CyberArk CIO Omer Grossman, Capsule gains credibility among security leaders who are grappling with rapid AI adoption. By providing real‑time governance, the startup enables enterprises to deploy agents faster without sacrificing compliance or risk posture. The move also signals a broader market trend: investors and incumbents are recognizing AI‑specific security as a distinct vertical, likely spurring additional funding and the emergence of standards for agentic runtime protection.