CodeStylist gives novice programmers course-specific code style feedback

CodeStylist, described in a new paper accepted at Frontiers in Education 2026, targets a stubborn problem in early programming courses: students are expected to follow local conventions for naming, formatting, comments, organization, and readability, yet feedback on those implicit standards is often delayed or inconsistent. The web application, built by researchers Ethan Dickey, Libra Vento, Peter Kurto, and Andres Bejarano, goes beyond generic linters and general-purpose LLM prompts by embedding course-specific standards directly into its feedback loop. It supports multi-file submissions and produces file-local and line-local explanations, aiming to guide revisions rather than simply score correctness.

To evaluate the prototype, the researchers ran a formative expert review with 18 instructional staff from one early undergraduate programming course. Participants explored the tool with self-selected code artifacts and then rated perceived correctness at 60.7%, response helpfulness at 3.50/5, usefulness at 3.33/5, and anticipated student learning at just 2.61/5. Despite those modest scores, 17 of 18 respondents expected students to use the tool primarily for style checking, often at least weekly. Open-ended feedback highlighted that CodeStylist made implicit course standards more visible, but instructors worried about unreliable output, student overreliance, and latency or cost. The authors conclude that course-aware style feedback is promising as a pre-submission aid, and recommend combining deterministic rule checks with LLM-generated explanations, rule citations, and stronger verification support in future versions.