Kodus

**名称:** Kodus **描述:** Kodus is an open source code review tool powered by artificial intelligence. It is designed to integrate directly into development workflows, enabling teams to automate the inspection of pull requests for quality, security, and performance issues. Unlike many proprietary code review services, Kodus gives users full control over which AI model they choose to run reviews, as well as the associated costs, because it operates on a model‑agnostic, zero‑markup pricing model. This means that a team can select a model from providers like OpenAI, Anthropic, or local open source models without being locked into a single vendor or paying inflated per‑review fees. Kodus learns from a team’s specific coding patterns and review history. Over time, it adapts its suggestions to align with the conventions and practices that the team already uses, reducing noise and false positives. The tool performs checks across multiple dimensions: code correctness, potential security vulnerabilities (such as injection flaws or hard‑coded secrets), performance bottlenecks, and adherence to style guides. It also scans for logical errors, dead code, and anti‑patterns that could lead to maintenance difficulties. The platform offers a generous free tier that allows teams to evaluate the service without a credit card. For those who need more capacity, a 14‑day free trial of the paid plan is available. Kodus can be triggered from multiple entry points: directly from a Git repository (via webhooks or a command‑line interface) or through a terminal session, making it suitable for teams using GitHub, GitLab, Bitbucket, or even self‑managed repositories. The tool’s configuration is highly flexible; users can define custom rules, adjust the severity thresholds for different kinds of feedbac