# Talk Freely, Execute Strictly: Schema-Gated Agentic AI for Flexible and Reproducible Scientific Workflows

arXiv: https://arxiv.org/abs/2603.06394 · HTML: https://mlsanity.com/paper/2603.06394/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 19 of its 63 references that are in the map.
- First posted on arXiv 6 Mar 2026.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 19 of 19)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Survey of Hallucination in Natural Language Generation](https://mlsanity.com/paper/2202.03629/) (2022)
- [ChemCrow: Augmenting large-language models with chemistry tools](https://mlsanity.com/paper/2304.05376/) (2023)
- [Executable Code Actions Elicit Better LLM Agents](https://mlsanity.com/paper/2402.01030/) (2024)
- [Defending Against Indirect Prompt Injection Attacks With Spotlighting](https://mlsanity.com/paper/2403.14720/) (2024)
- [Security and Privacy Challenges of Large Language Models: A Survey](https://mlsanity.com/paper/2402.00888/) (2024)
- [The Task Shield: Enforcing Task Alignment to Defend Against Indirect Prompt Injection in LLM Agents](https://mlsanity.com/paper/2412.16682/) (2024)
- [An introduction to Docker for reproducible research, with examples from the R environment](https://mlsanity.com/paper/1410.0846/) (2014)
- [A Computational Inflection for Scientific Discovery](https://mlsanity.com/paper/2205.02007/) (2022)
- [Unnatural language processing: How do language models handle machine-generated prompts?](https://mlsanity.com/paper/2310.15829/) (2023)
- [The (R)evolution of Scientific Workflows in the Agentic AI Era: Towards Autonomous Science](https://mlsanity.com/paper/2509.09915/) (2025)
- [WorkflowHub: a registry for computational workflows](https://mlsanity.com/paper/2410.06941/) (2024)
- [Do Large Language Models Speak Scientific Workflows?](https://mlsanity.com/paper/2412.10606/) (2024)
- [Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition](https://mlsanity.com/paper/2510.15280/) (2025)
- [Design of a nickel-base superalloy using a neural network](https://mlsanity.com/paper/1803.03039/) (2018)
- [Ensuring Reproducibility in Generative AI Systems for General Use Cases: A Framework for Regression Testing and Open Datasets](https://mlsanity.com/paper/2505.02854/) (2025)
- [From Prompt to Pipeline: Large Language Models for Scientific Workflow Development in Bioinformatics](https://mlsanity.com/paper/2507.20122/) (2025)
- [Unveiling Large Language Model Supply Chain: Structure, Domain, and Vulnerabilities](https://mlsanity.com/paper/2504.20763/) (2025)
- [Report of the 2025 Workshop on Next-Generation Ecosystems for Scientific Computing: Harnessing Community, Software, and AI for Cross-Disciplinary Team Science](https://mlsanity.com/paper/2510.03413/) (2025)

## Built on by (most cited first, 1 of 1)

- [It's not the Language Model, it's the Tool: Deterministic Mediation for Scientific Workflows](https://mlsanity.com/paper/2605.13245/) (2026)

## Abstract (arXiv)

Large language models (LLMs) can now translate a researcher's plain-language goal into executable computation, yet scientific workflows demand determinism, provenance, and governance that are difficult to guarantee when an LLM decides what runs. Semi-structured interviews with 18 experts across 10 industrial R&D stakeholders surface 2 competing requirements--deterministic, constrained execution and conversational flexibility without workflow rigidity--together with boundary properties (human-in-the-loop control and transparency) that any resolution must satisfy. We propose schema-gated orchestration as the resolving principle: the schema becomes a mandatory execution boundary at the composed-workflow level, so that nothing runs unless the complete action--including cross-step dependencies--validates against a machine-checkable specification. We operationalize the 2 requirements as execution determinism (ED) and conversational flexibility (CF), and use these axes to review 20 systems spanning 5 architectural groups along a validation-scope spectrum. Scores are assigned via a multi-model protocol--15 independent sessions across 3 LLM families--yielding substantial-to-near-perfect inter-model agreement (Krippendorff a=0.80 for ED and a=0.98 for CF), demonstrating that multi-model LLM scoring can serve as a reusable alternative to human expert panels for architectural assessment. The resulting landscape reveals an empirical Pareto front--no reviewed system achieves both high flexibility and high determinism--but a convergence zone emerges between the generative and workflow-centric extremes. We argue that a schema-gated architecture, separating conversational from execution authority, is positioned to decouple this trade-off, and distill 3 operational principles--clarification-before-execution, constrained plan-act orchestration, and tool-to-workflow-level gating--to guide adoption.