answer-processing

To grade scanned handwritten PDF answers, this Skill converts the submissions into markdown using OCR, then applies strategy-based comparison against your reference solution to evaluate reasoning patterns and completeness. It returns a compact grading verdict with structured error logs.

Yes, you can grade quiz answers with noisy or unavailable OCR because the Skill applies strategy-based grading that evaluates reasoning patterns and variable choices instead of demanding perfect text recognition accuracy. It also handles directly pasted student work.

The best way to convert scanned homework PDFs into markdown for grading is using the OCR-aware conversion process built into this Skill, which cleans the extracted text and compares it directly against answer keys or twin problems.

The grading strategy checks for both the final end form and the overall answer by evaluating reasoning patterns, variable choices, and completeness, ensuring students receive credit for correct methodology even when OCR results are imperfect.

Error logging works by appending structured grading issues to a canonical log during the evaluation of handwritten answer PDFs, allowing you to review OCR-related notes and specific grading discrepancies through a returned compact summary.