Intent-Driven Computing: A Computational Model for Governed Autonomous Systems

This collection contains multiple CS research papers with clear security and supply-chain implications. "Intent-Driven Computing" proposes a model where all effects are issued as finite intents checked by a decidable policy language and recorded in a tamper-evident ledger — improving auditability, forensics, and governance but creating high-value policy enforcement/runtime components that, if flawed or compromised, could become single points of failure or escalation vectors. "Context" and related work on composable sandboxed LM-generated programs and proactive agents advance sandboxing and-gov