Prevalence study of 10k GitHub repos identifies duplication, low revision, and missing permissions in LLM agent configs; proposes Rel(AI)Build control plane with content addressing, tiered permissions, state machine, and Jaccard drift detection.
Nemo guardrails: A toolkit for controllable and safe llm applications with programmable rails
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Deterministic Control Plane for LLM Coding Agents
Prevalence study of 10k GitHub repos identifies duplication, low revision, and missing permissions in LLM agent configs; proposes Rel(AI)Build control plane with content addressing, tiered permissions, state machine, and Jaccard drift detection.