Research
Research July 2026
Haurx Agent: transforming engineering work
A research note on Haurx Agent's controlled engineering work cycle and Vtslx AO collaboration.15 min read →Vauix Agent: transforming agentic work
A research note on Vauix Agent's local execution and governed collaboration for complex work.
14 min read →Vtslx AO: sensitive context handling
A public architecture and governance note for minimizing, routing, retaining, and auditing highly sensitive AI context.
19 min read →Vtslx AO LLM security and three-layer interception policy
A public research note on three-layer LLM security interception, tool authorization, streaming termination, and safe policy semantics.
20 min read →Vtslx AO cache architecture and semantics
A public research note on safe cache eligibility, invalidation, streaming, tool semantics, and observable fallbacks for an AI gateway.
18 min read →Orchestrations v1: a public architecture for observable model-tool runs
A public architecture and measurement note for a gateway-native streaming model-and-tool orchestration protocol.
22 min read →Civilization Architecture: auditable Memory, State, and Rule pathways
A research note on integrating structured Memory, State, and Rule pathways into hidden-state computation with explicit ablation and audit surfaces.
48 min read →Privacy and sandboxing for verifiable AI workflows
A local-first, verifiable architecture for privacy, sandboxing, AI context, and agent tools.
32 min read →Hig: project-aware incremental archiving for developer workflows
How Hig coordinates change awareness, reuse, bounded memory, and a stable archive surface for recurring developer workflows.
8 min read →Context Hub: give AI a map of your entire project
Why a local semantic index gives AI a durable map of a codebase—without treating your repository as a prompt attachment.
18 min read →