Building infrastructure for production AI agent systems.
We're focused on the gap between prototype and production: helping engineering teams ship agents that are reliable, verifiable, and cost-efficient in the real world.
Current project: MCP Firewall for Agents
Follow along or reach out if you're working on production agent systems.
Policy-as-code enforcement and observability for MCP tool calls. Wraps AI agent sessions with cryptographic integrity checks, argument-level CEL policies, and a full audit trail.
Security proxy for MCP servers. Enforces policy-as-code on every tool call between AI agent runtimes like Claude Cowork and upstream MCP servers. Powered by the Deconvolute SDK.
Live PoC: MCP attacks that compromise AI agents mid-session and how to block them in a few lines of code.
Reproducible security benchmarking for the Deconvolute SDK and AI system integrity against adversarial attacks.
Automatically generate YARA rules from adversarial and benign text samples. Built for detecting indirect prompt injection attacks on RAG pipelines.
A demonstration of RAG poisoning attacks using dormant documentation injections.
Loading…
Loading…
