Problem
Research questions and friction points this paper is trying to address.
agent skills
control flow
task reasoning
finite state machines
skill compilation
HEXIS: Compiling Skills into Extended Finite State Machines
🤖 AI Summary
This study addresses the problem of omitted or misapplied steps caused by the coupling of reasoning and control in agent skill execution. To decouple domain knowledge from control flow, this work proposes compiling skills into Extended Finite State Machines (EFSMs). The core methodology involves designing an incremental compiler that integrates static analysis with a trajectory replay algorithm, enabling automatic alignment between skills and EFSMs, tool interface mapping, and dynamic updates. Experimental results demonstrate that the proposed approach improves average task success rates by 16.1% while reducing execution token consumption by 38.4% to 88.9%, significantly enhancing both the reliability and efficiency of agent execution.
📝 Abstract
Agent skills provide reusable knowledge and instructions, yet agents must repeatedly infer how to apply them and which operation should follow. This couples task reasoning with control decisions, allowing prescribed steps to be omitted or applied incorrectly. We introduce HEXIS, which compiles agent skills into extended finite state machines that separate knowledge from control flow. Skill knowledge is incorporated into local instructions that guide reasoning and generation within states. The machine records execution progress and intermediate results, while explicit transition conditions determine subsequent operations. Our incremental compiler first maps skill clauses and tool interfaces to state operations, local instructions, data bindings, and transitions. It then aligns development traces with existing states to identify missing operations and dependencies. These are incorporated by adding or reusing states and refining their connections. Updates are accepted only after static checks and replay of the current and all previously accepted traces. Across four benchmarks and four executors, HEXIS improves success over Skill + ReAct by 16.1 percentage points on average. Qwen3.8-27B reduces execution tokens by 38.4-88.9% across benchmarks.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Extended Finite State Machines
Skill Compilation
Incremental Compiler
Control Flow Decoupling
Replay Verification
🔎 Similar Papers
No similar papers found.
