Atsign Launches AI Architect to Add Security Controls to AI-Built Apps
Summary
Atsign says its AI Architect product uses a custom MCP server, cryptographic identities and policy controls to guide AI coding agents in building applications. The company claims its approach keeps credentials and resources hidden from attackers.
Key points
- AI Architect helps developers define an application blueprint and generates JSON-based prompts with build instructions and SDK references for a chosen coding agent.
- The product uses a custom MCP server, called AAIA, to apply authentication, authorization, encryption and context-based policies to interactions among resources.
- Atsign says each resource receives a unique cryptographic identity with permissions controlled by policies.
- The company says cryptographic keys are non-custodial and its relay servers store ciphertext rather than cleartext or credentials.
- Atsign claims its approach leaves no open ports or public APIs and makes resources invisible to attackers; the article provides no independent validation of these claims.
Article Details
- Event Type
- Atsign product launch: AI Architect, a platform for guiding AI-assisted application development with security policies and cryptographic identities.
- Impact
- Atsign says AI Architect generates prompts and security rules intended to help developers build applications with authenticated, authorized, encrypted, and governed resource interactions. The company claims its approach hides identities and credentials from attackers and leaves no open ports or public APIs; the article reports no independently verified security outcomes or specific affected organizations.