# Reinforced Agent: Inference-Time Feedback for Tool-Calling Agents

arXiv: https://arxiv.org/abs/2604.27233 · HTML: https://mlsanity.com/paper/2604.27233/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use.
- Builds on 6 of its 10 references that are in the map.
- First posted on arXiv 29 Apr 2026.

## Placed in

- LLM agents and tool use / [Tool learning and function calling](https://mlsanity.com/agents/tool-learning-and-function-calling/)

## Builds on (most cited first, 6 of 6)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Gorilla: Large Language Model Connected with Massive APIs](https://mlsanity.com/paper/2305.15334/) (2023)
- [Large Language Models Are Reasoning Teachers](https://mlsanity.com/paper/2212.10071/) (2022)
- [Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programs](https://mlsanity.com/paper/2406.11695/) (2024)
- [Specializing Smaller Language Models towards Multi-Step Reasoning](https://mlsanity.com/paper/2301.12726/) (2023)

## Built on by (most cited first, 7 of 7)

- [Verifier-Guided Code Translation via Meta-Step Decoding](https://mlsanity.com/paper/2605.17626/) (2026)
- [Outcome Monitors: Recovery Affordances for Silent Tool Failures](https://mlsanity.com/paper/2608.19303/) (2026)
- [BenchShield: Formal Model-Backed Instrumentation for Reward Integrity in LLM-Agent Evaluation Infrastructure](https://mlsanity.com/paper/2609.11028/) (2026)
- [Runtime Action Interference for AI Control of AlphaStar in StarCraft II](https://mlsanity.com/paper/2608.21398/) (2026)
- [Dissecting Agentic Forensics: The Role of Triage, Prompting, and Evidence Arbitration in Open-World Fake Image Detection](https://mlsanity.com/paper/2609.24359/) (2026)
- [What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agents](https://mlsanity.com/paper/2609.39989/) (2026)
- [RSure-Agent: Reliable Use of Tool Observations for Remote Sensing Agents](https://mlsanity.com/paper/2610.04836/) (2026)

## Abstract (arXiv)

Tool-calling agents are evaluated on tool selection, parameter accuracy, and scope recognition, yet LLM trajectory assessments remain inherently post-hoc. Disconnected from the active execution loop, such assessments identify errors that are usually addressed through prompt-tuning or retraining, and fundamentally cannot course-correct the agent in real time. To close this gap, we move evaluation into the execution loop at inference time: a specialized reviewer agent evaluates provisional tool calls prior to execution, shifting the paradigm from post-hoc recovery to proactive evaluation and error mitigation. In practice, this architecture establishes a clear separation of concerns between the primary execution agent and a secondary review agent. As with any multi-agent system, the reviewer can introduce new errors while correcting others, yet no prior work to our knowledge has systematically measured this tradeoff. To quantify this tradeoff, we introduce Helpfulness-Harmfulness metrics: helpfulness measures the percentage of base agent errors that feedback corrects; harmfulness measures the percentage of correct responses that feedback degrades. These metrics directly inform reviewer design by revealing whether a given model or prompt provides net positive value. We evaluate our approach on BFCL (single-turn) and Tau2-Bench (multi-turn stateful scenarios), achieving +5.5% on irrelevance detection and +7.1% on multi-turn tasks. Our metrics reveal that reviewer model choice is critical: the reasoning model o3-mini achieves a 3:1 benefit-to-risk ratio versus 2.1:1 for GPT-4o. Automated prompt optimization via GEPA provides an additional +1.5-2.8%. Together, these results demonstrate a core advantage of separating execution and review: the reviewer can be systematically improved through model selection and prompt optimization, without retraining the base agent.