mlsanityData to 7 Oct 2026

JTPRO: A Joint Tool-Prompt Reflective Optimization Framework for Language Agents

arXiv 2604.19821 ↗First posted 20 Apr 2026cs.AI cs.SEcited by 4

On the map: LLM agents and tool use → Tool learning and function calling

Abstract (arXiv)

Large language model (LLM) agents augmented with external tools often struggle as number of tools grow large and become domain-specific. In such settings, ambiguous tool descriptions and under-specified agent instructions frequently lead to tool mis-selection and incorrect slot/value instantiation. We hypothesize that this is due to two root causes: generic, one-size-fits-all prompts that ignore tool-specific nuances, and underspecified tool schemas that lack clear guidance on when and how to use each tool and how to format its parameters. We introduce Joint Tool-Prompt Reflective Optimization (JTPRO), a framework for improving tool-calling reliability in trace-supervised settings by iteratively using rollout-driven reflection to co-optimize global instructions and per-tool schema/argument descriptions for accurate tool selection and argument instantiation in large tool inventories. JTPRO is designed to preserve only tool-local cues needed for correct disambiguation and slot filling. We evaluate JTPRO across multi-tool benchmarks, which account for different number of tools using three metrics: Tool Selection Accuracy (TSA), Slot Filling Accuracy(SFA), and Overall Success Rate(OSR) (correct tool + correct slots + correct values). JTPRO consistently outperforms strong baselines, including CoT-style agents, and reflective prompt optimizers such as GEPA by 5%-20% (relative) on OSR. Ablations show that joint optimization of instructions and tool schemas is more effective and robust than optimizing either component in isolation.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use. Builds on 30 of its 47 references that are in the map. First posted on arXiv 20 Apr 2026.

Builds on

30 of 47 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Procedural Environment Generation for Tool-Use Agents21 May
  2. 2024Meta-Reasoning Improves Tool Use in Large Language Models7 Nov
  3. 2024CITI: Enhancing Tool Utilizing Ability in Large Language Models without Sacrificing General Performance20 Sep
  4. 2024Seal-Tools: Self-Instruct Tool Learning Dataset for Agent Tuning and Detailed Benchmark14 May
  5. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs31 Jul
  6. 2023Toolformer: Language Models Can Teach Themselves to Use Tools9 Feb

Other references · 24

  1. 2017Attention Is All You Needcited by 59,932
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2021The Power of Scale for Parameter-Efficient Prompt Tuningcited by 3,769
  4. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
  5. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,203
  6. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 803
  7. 2023Automatic Prompt Optimization with "Gradient Descent" and Beam Searchcited by 628
  8. 2025GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningcited by 449
  9. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 377
  10. 2025ToolRL: Reward is All Tool Learning Needscited by 342
  11. 2024Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programscited by 270
  12. 2023ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Modelscited by 174
12 more
  1. 2023RestGPT: Connecting Large Language Models with Real-World RESTful APIscited by 129
  2. 2023CRAFT: Customizing LLMs by Creating and Retrieving from Specialized Toolsetscited by 122
  3. 2024AvaTaR: Optimizing LLM Agents for Tool Usage via Contrastive Reasoningcited by 69
  4. 2025Feedback Descent: Open-Ended Text Optimization via Pairwise Comparisoncited by 42
  5. 2024ToolRerank: Adaptive and Hierarchy-Aware Reranking for Tool Retrievalcited by 40
  6. 2025AutoPDL: Automatic Prompt Optimization for LLM Agentscited by 27
  7. 2025Hard Negative Mining for Domain-Specific Retrieval in Enterprise Systemscited by 14
  8. 2025Aligning LLMs for Multilingual Consistency in Enterprise Applicationscited by 10
  9. 2025Maestro: Joint Graph & Config Optimization for Reliable AI Agentscited by 9
  10. 2025Can LLMs Narrate Tabular Data? An Evaluation Framework for Natural Language Representations of Text-to-SQL System Outputscited by 7
  11. 2025Hybrid AI for Responsive Multi-Turn Online Conversations with Novel Dynamic Routing and Feedback Adaptationcited by 6
  12. 2026DiffuMask: Diffusion Language Model for Token-level Prompt Pruningcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 4.

4 papers.

  1. 2026SkillsInjector: Dynamic Skill Context Construction for LLM Agentscited by 5
  2. 2026MemeMind: Reference-Guided Trace Construction for Offline Context Optimizationcited by 0
  3. 2026DOCSCHISEL: Adaptive Tool Documentation Optimization Framework for LLM Agentscited by 0
  4. 2026Adaptive-GEPA: Make Your Harness Fit Heterogeneous Requestscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026ToolAtlas: Learning Once, Reusing Everywhere with Tool-Side Memory
  2. 2026ToolChoiceConfusion: Causal Minimal Tool Filtering for Reliable LLM Agents
  3. 2026Joint Optimization of Tool Creation and Use for Large Language Model Agents
  4. 2026ToolMenuBench: Benchmarking Tool-Menu Filtering Strategies for Reliable and Efficient LLM Agents
  5. 2025Advancing Tool-Augmented Large Language Models via Meta-Verification and Reflection Learning
  6. 2026Tools Are Not Islands: Set-Level Tool Retrieval for LLM Agents via Query-Conditioned Hyperedge Prediction
  7. 2025ToolScope: Enhancing LLM Agent Tool Use through Tool Merging and Context-Aware Filtering
  8. 2026ToolSearcher: Optimizing Tool Selection at Scale via Reinforcement Learning