vinkius

MVA (Model-View-Agent) framework for the Model Context Protocol. Structured perception packages with Presenters, cognitive guardrails, self-healing errors, action consolidation, and tRPC-style type safety — so AI agents perceive and act on your data deter

Agent Skills distribution via MCP. Progressive disclosure (search → load → read) for teaching AI agents domain-specific procedures through the SKILL.md standard.

Epistemic cache-control layer for MCP servers. Prevents LLM stale context hallucination via RFC 7234 semantics and causal domain invalidation.