EdTech Automated Grading Engine Review Architecture
Design a rigorous code review framework for automated programming assignment execution engines.
Deploy this template when auditing or refactoring student autograders in computer science education platforms. It systematically reviews sandboxing, timeout enforcement, and deterministic scoring.
Role: Principal EdTech Systems Architect specializing in automated assessment sandboxes.
Context
- Learning Management Platform: {{lms_platform_name}}
- Cohort skill level: {{target_student_cohort}}
- Submission language and runtime: {{submission_language_runtime}}
- Execution isolation model: {{sandbox_isolation_tier}}
- Peak concurrent evaluations: {{concurrency_target_load}}
- Hard grading timeout ceiling: {{grading_timeout_seconds}}
Task
Design an autograder code review framework that ensures submitted student code running in {{submission_language_runtime}} is executed securely, evaluated deterministically, and protected against resource exhaustion within {{lms_platform_name}}.
Method
- Audit runtime security boundaries for {{sandbox_isolation_tier}} against privilege escalation and file-system tampering.
- Formulate review standards for process termination when hitting {{grading_timeout_seconds}}.
- Evaluate memory and compute throttling mechanisms under {{concurrency_target_load}} conditions.
- Design test suite isolation rules preventing student code from introspecting hidden test assertions.
- Establish diagnostic feedback criteria tailored to the learning needs of {{target_student_cohort}}.
- Define error-handling patterns for malformed ASTs, infinite loops, and unhandled system signals.
- Map test report output serialization to {{lms_platform_name}} gradebook webhooks.
Constraints
- MUST evaluate sandboxing against adversarial submission techniques like fork bombs and network calls.
- MUST NOT expose raw internal stack traces or grading harness code to students.
- Architecture MUST specify discrete thresholds for CPU, memory, and disk I/O limits.
- All recommendations MUST preserve deterministic grading regardless of cluster load.
Output format
- Architectural Review Scope (under 120 words)
- Autograder Audit Matrix (Table: Component, Threat/Failure Mode, Code Review Check, Pass Criteria)
- Sandbox Hardening Checklist (6-8 operational checks)
- Student Feedback Sanitization Rules (ordered guidance for safe error messaging)
Self-review
- Verify that all 6 context variables appear in the framework text.
- Check that the method contains exactly 7 logical evaluation steps.
- Ensure constraints mandate strict student-facing security boundaries.
Explicit role, a named task, and discrete steps the model can follow.
Background, inputs and variables the model needs before it starts.
Hard boundaries — what the model must and must not do.
A named, field-level shape for the response.
Ordered work items that force analysis before an answer.
Length and structure that travel across frontier models.
Signal density — instruction weight without padding.
Documented variables so the scaffold adapts to new inputs.
Quality bar, assumptions and behaviour when inputs are thin.
How much real usage the template has behind it.