Task Decomposition-Guided Reranking for Adaptive Agent Skill Retrieval

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Yanping Chen, Weijie Shi, Wen Yang, Jiajie Xu
• Venue: arXiv: cs.AI
• Published: 2026-07-07
• Collected: 2026-07-08
• URL: http://arxiv.org/abs/2607.06283v1
• PDF: https://arxiv.org/pdf/2607.06283v1
• DOI: UNKNOWN

Plain-Language Brief

The authors propose a method called SkillReranker to improve the accuracy of selecting skills for complex tasks in agent systems. This is achieved by semantically decomposing both tasks and skills, creating a structured task-skill correspondence, identifying subtask intervals, and using a cross-encoder to score candidate skills. The method improves task performance, reduces environment interaction steps, and lowers token consumption compared to existing methods.

What the Document Says

• The authors propose SkillReranker for adaptive skill selection.
• SkillReranker uses semantic decomposition on tasks and skills.
• It constructs a directed acyclic execution graph with task states as nodes and skills as edges.
• For each task interval, it employs a cross-encoder to score candidate skills.
• The method improves task performance, reduces environment interaction steps, and lowers token consumption.

Key Concepts Mentioned

• SkillReranker
• semantic decomposition
• directed acyclic execution graph
• task interval
• cross-encoder

Claims Made by the Document

• Existing methods overlook the dynamic influence of task difficulty and skill applicability.
• SkillReranker effectively improves task performance.

Evidence Used

• The authors mention that existing methods tend to overlook the dynamic influence of task difficulty and skill applicability.
• The results from experiments show improvement in task performance.

Methods / Approach

• Semantic decomposition
• Construction of a directed acyclic execution graph
• Cross-encoder for scoring candidate skills

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Skill usage can significantly enhance the ability of modern agent systems to complete complex tasks. However, the growing scale of skill libraries makes accurate skill selection increasingly challenging. In real-world scenarios, ambiguous semantic matching often arises between a specific task requirement and multiple generic yet semantically similar candidate skills. Moreover, existing methods tend to overlook the dynamic influence of task difficulty and skill applicability when selecting the optimal target skill set. To address these issues, we propose SkillReranker, an inference-time reranking framework for adaptive skill selection. Specifically, we first perform semantic decomposition on both the task and skill sides, yielding informative subtask and execution-state descriptions as well as transition-state descriptions that characterize each skill's functionality. These descriptions are then used to construct a directed acyclic execution graph, where intermediate task states are modeled as nodes and candidate skills as edges, thereby establishing a structured task-skill correspondence. On this basis, SkillReranker determines whether each state node satisfies the split condition to identify subtask intervals. For each task interval, we employ a cross-encoder to perform comprehensive scoring over candidate skills and select the most suitable ones to form the final target skill set. Experiments on ALFWorld and ScienceWorld with three backbone LLMs show that SkillReranker effectively improves task performance, reduces environment interaction steps, and lowers token consumption compared with existing skill selection baselines.