PUBLICATIONS

We publish what we find, including what doesn't work. The field moves faster when the work is open.

1Publications
100%Open Access
Rasin-related methodsMultilingual Archival Retrieval

Cross-lingual search, citation verification, entity graphs, and claim provenance for low-resource and postcolonial archives. Being developed through Rasin, with reusable methods published as they mature.

studio1804 lab researchSecurity of Autonomous AI Infrastructure

Attack surfaces at the agent-tool boundary when autonomous systems invoke external tools and services. Focus: infrastructure-level isolation and capability containment.

AI SAFETYAI SECURITYAUTONOMOUS AGENTS2026

The Capability-Container Pattern: Infrastructure-Level Security for Autonomous AI Agents

Ricardo Ledan

Addresses security vulnerabilities at the agent-tool boundary where autonomous AI agents invoke external tools via protocols like MCP. Proposes an infrastructure-level approach where agents never directly access tools—all tool invocations flow through a mediation gateway into isolated containers with minimal provisioned capabilities.

ZenodoDOI: 10.5281/zenodo.18614503READ PAPER →