What problem does it solve?
Teachers, TAs, and mentors often spend hours manually checking student code against assignment requirements, struggling to produce evidence-backed feedback, avoid exposing hidden grading details to students, and handle edge cases like missing design documents or broken submission environments. This skill automates the entire audit process with strict, design-first principles to ensure consistent, fair grading.
Core Features & Use Cases
- Design-first compliance checking: Trace every assignment requirement to concrete code evidence, rather than relying on general style preferences.
- Evidence-backed findings: Cite exact file paths and line numbers for all pass, fail, and drift results.
- Safe feedback separation: Automatically split student-facing feedback from private instructor notes to avoid leaking hidden rubrics or grading shortcuts.
- Use Case: A TA grading 50 machine learning project submissions can use this skill to quickly verify each submission follows the required Hydra configuration structure, API design, and test coverage, while keeping all hidden grading criteria confidential.
Quick Start
Use the student-codebase-audit skill to review the attached student repository against the provided assignment rubric and generate a full audit report with findings and feedback.