OctoBench: Benchmarking Scaffold-Aware Instruction Following in Repository-Grounded Agentic Coding
This addresses the need for better benchmarking of instruction-following in coding agents, which is incremental but important for reproducible development.
The paper tackles the problem of evaluating how well LLM-based coding agents follow scaffold-specified instructions in repository-grounded tasks, introducing OctoBench with 34 environments and 217 tasks across three scaffold types, and finds a systematic gap between task-solving and scaffold-aware compliance across eight models.
Modern coding scaffolds turn LLMs into capable software agents, but their ability to follow scaffold-specified instructions remains under-examined, especially when constraints are heterogeneous and persist across interactions. To fill this gap, we introduce OctoBench, which benchmarks scaffold-aware instruction following in repository-grounded agentic coding. OctoBench includes 34 environments and 217 tasks instantiated under three scaffold types, and is paired with 7,098 objective checklist items. To disentangle solving the task from following the rules, we provide an automated observation-and-scoring toolkit that captures full trajectories and performs fine-grained checks. Experiments on eight representative models reveal a systematic gap between task-solving and scaffold-aware compliance, underscoring the need for training and evaluation that explicitly targets heterogeneous instruction following. We release the benchmark to support reproducible benchmarking and to accelerate the development of more scaffold-aware coding agents.