Invariant Labs

@InvariantLabsAI

Invariant Labs makes AI Agents secure and reliable.

Joined July 2024

28Following

653Followers

Invariant Labs Retweeted

Luca Beurer-Kellner@lbeurerkellner · May 26

😈 BEWARE: Claude 4 + GitHub MCP will leak your private GitHub repositories, no questions asked. We discovered a new attack on agents using GitHub’s official MCP server, which can be exploited by attackers to access your private repositories. creds to @marco_milanta (1/n) 👇

494

2.0K

499.0K

Invariant Labs@InvariantLabsAI · May 26

Invariant researchers have uncovered a new security flaw in GitHub’s official MCP server, enabling attackers to exfiltrate private repository data. The toxic flow was identified during an automated scan using Invariant's security stack. Learn more: invariantlabs.ai/blog/mcp-githu…

InvariantLabsAI's tweet card. We showcase a critical vulnerability with the official GitHub MCP server, allowing attackers to access private repository data. The vulnerability is among the first discovered by Invariant's security...

461

Invariant Labs@InvariantLabsAI · May 22

🔵 New release: Invariant MCP-scan v0.2 is here! Track, audit & secure all local MCP traffic with static+dynamic scanning, local guardrails, and customizable policies. Ideal for orgs prioritizing agent security & compliance. Docs: explorer.invariantlabs.ai/docs/mcp-scan/ #AI #DevSecOps

InvariantLabsAI's tweet image. 🔵 New release: Invariant MCP-scan v0.2 is here!

Track, audit &amp; secure all local MCP traffic with static+dynamic scanning, local guardrails, and customizable policies.

Ideal for orgs prioritizing agent security &amp; compliance.

Docs: explorer.invariantlabs.ai/docs/mcp-scan/

#AI #DevSecOps

283

Invariant Labs@InvariantLabsAI · Apr 29

We are proud to share that AgentDojo, an Invariant research project done with @ETH, has won the first price of the @ai_risks SafeBench competition. We truly appreciate this recognition from the community. Learn More: invariantlabs.ai/blog/agentdojo…

InvariantLabsAI's tweet card. AgentDojo wins $50,000 prize in the SafeBench competition, hosted by the prestigious Center for AI Safety.

2.0K

Invariant Labs Retweeted

Bill Doerrfeld@DoerrfeldBill · Apr 24

MCP is the hottest thing in AI right now, but people aren't really talking about the security implications... I covered a recently discovered exploit and mitigations on the @thenewstack today: thenewstack.io/building-with-…

2.0K

Invariant Labs Retweeted

Marc Fischer@marc_r_fischer · Apr 24

We recently shipped a lot of updates to mcp-scan: - whitelisting of tools - Improvements to the server (reducing false-positives, improving detection) - run via npm/npx Much more coming soon! github.com/invariantlabs-… #mcp

874

Invariant Labs Retweeted

Luca Beurer-Kellner@lbeurerkellner · Apr 15

🔴🌎 New MCP attack on BrowserMCP We show an MCP attack on the popular BrowserMCP. It allows attackers to read arbitrary files from your machine, when the agent visits the website below. Try yourself with: access.invariantlabs.ai No bad MCP server needed. (1/n)👇

31.0K