Working With Sleepers
Research.
Research-grade infrastructure and security middleware for selective organizational environments.
The laboratory occasionally partners with external organizations who want research-grade AI security systems implemented directly within their own infrastructure. These engagements typically focus on deploying customized instances of security middleware, such as the technologies developed under Project Black Monolith, or implementing custom grounding and evaluation frameworks.
Because we operate primarily as an independent research unit, these collaborations are highly selective. We do not provide standardized consulting services or fixed product packages. Every engagement is scoped individually from first principles to ensure that the work aligns with our core research interests in agentic safety, autonomous agent security, and security middleware.
Organizations interested in discussing a custom deployment, security audit, or technical partnership should reach out to the laboratory directly. We review inquiries on a case-by-case basis.
Custom Middleware Integration
Deploy tailored instances of MCP-Shield and VectorAnchor within your private infrastructure to enforce zero-trust tool execution policies.
Agent Safety & Security Audits
In-depth empirical evaluations of autonomous agent reason loops, tool schemas, and vulnerabilities against adversarial prompt vectors.