Code Review Assistant

Evaluation suite for the AI-powered code review tool. Focuses on suggestion accuracy and security vulnerability detection.

FAIL

Public

run-005

OpenAI

Auth: Server KeyModel: gpt-4o-miniHealth: Not Tested

Pass Rate Min85%

Faithfulness Min0.8

PII Max0

Secrets Max0

Jailbreak Max2

PIIMonitor

SecretsBlocking

JailbreakBlocking

ToxicityMonitor

Case Uploads

Status	File	Template	Cases	Date
SUCCESS	code_review_new_cases.csv	Standard Eval Cases	32	Dec 11, 2025

Result	Run ID	Trigger	Model	Execution	Pass Rate	Accuracy Delta	Started
FAIL	run-005	PR	gpt-4o-mini	Simulated	75%	-8.5%	10h ago
WARN	run-006	Nightly	gpt-4o	Simulated	84.4%	-3.1%	1d ago