Qodo 用 Codex 将上下文构建为 AI 代码审查
Qodo 带来组织上下文和代码库标准进入Codex，帮助开发者在寄存器之间捕捉P0问题。
Qodo(在新窗口打开)为代码检讨和软件质量构建AI工具，给agents提供理解他们更改代码所需的组织知识。 共同创始人兼首席执行官伊塔玛尔·弗里德曼(Itamar Friedman)称其为"编码agent最好的朋友".
Qodo使用OpenAI模型包括GPT‐6 Astra来收集工程知识，审查代码，并监测软件质量。 其代码插件带来上下文并审查开发者的编码会话。
“我们爱OpenAI模型。 我们想要前沿，最好的推理模式，但也想要真正快速的推理模式。”
Itamar描述了可以跨越10,000或更多寄存器的客户环境，本地代码的更改可以打破下游服务，即使它自己看起来是正确的。 Qodo在 Codex 市场中建立了一个插件来解决这个问题。
Itamar 称。 “Codex非常善于承担当地的任务，Qodo的插件通过Qodo的代理工具箱，让 Codex 访问代码库知识，组织规则，以及独立审查。
在上下文中，Codex可以说明整个寄存器的依赖性，包括由不同的Git提供商托管或通过Kafka连接的服务。 “Qodo将进行这种连接，并将帮助Codex将软件作为一个整体看待，而不是错过这些P0bugs,”Itamar 称。
在 Codex 开始写入代码之前，它可以使用 Qodo 的插件检索相关的代码库规则并拉动请求的历史和模式。 在 Codex 进行修改后，Qodo 在开发者打开拉动请求之前，在同一个会话中将其结果直接返回 Codex.
Nnenna Ndukwe,Qodo的AI开发者关系领跑者，看到开发者使用这个工作流程来构建功能，并在Codex中用GPT‐6模型解决技术债务。 他们正在将Codex的编码工作与Qodo的独立审查之间的交接自动化。 “这就是我所看到的开发者所做的
Itamar将编码和审查agents描述为“阳和阳”：一个创建代码，另一个挑战代码。 每个角色都需要自己的设置和上下文，可能涉及多个agents. 他的愿景是让agents讨论调查结果，开发者将成为规划者和审查者，能够与agents的讨论、背景和事先开发者的决定一起检查未决问题。
”伊塔马尔·称。 Nnna 新增。 “我们可以利用我们专门知识的最佳部分。”
Nnenna在Qodo内部
“Codex没有做什么？我喜欢Codex所做的一切我喜欢将AI和自动化结合起来的想法。”
Nnenna安排 Codex 报告她在一周内的成就以及她接下来可以集中或改进的地方。 她还用它来研究社交平台上的代理工程，并提出开发者内容的想法。 在她自己的代码更改方面，她安排Codex来分析自己的作品，并建议agent技能来创建或精炼。
Itamar认为有机会使用OpenAI最近发布的发布决定API来选择哪些规则和agent技能适用于每个发展阶段，而不是依靠语言模型来决定每个决定。
他举了一个例子，即 " 决定API " 可以选择200条规则中的哪条规则需要通过相关规则在语言模式前检查，帮助加快选择速度和费用，同时提高审查质量。 “决定API为我们带来更大的信任，”Itamar 称。
Quodo的方法结合了"人工智能的最好方面和传统软件工程和脚本的最好方面，"Nnna 称。 公司知识指导编码agents，独立审查检查他们的工作，人的专门知识解决遗留的问题。
这为Itamar所称的软件厂奠定了基础：一个系统客户可以组装和拥有，将编码和审查agents，同时理解软件如何改变生产。
OpenAI 的Qodo 正在构建未来，将agents编码后
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Qodo builds context into AI code review with Codex
Qodo brings organizational context and codebase standards into Codex, helping developers catch P0 issues across repositories.
Qodo(opens in a new window) builds AI tools for code review and software quality, giving coding agents the organizational knowledge they need to understand the code they change. Co-founder and CEO Itamar Friedman calls it “the coding agent’s best friend.”
Qodo uses OpenAI models including GPT‑6 Astra to gather engineering knowledge, review code, and monitor software quality. Its Codex plugin brings that context and review into developers’ coding sessions.
“We love OpenAI models. We want the frontier, the best reasoning models, but also those that are really fast.”
Itamar describes customer environments that can span 10,000 or more repositories, where a local code change can break a downstream service, even if it looks correct on its own. Qodo built a plugin in the Codex marketplace to address this.
“Codex is really good at taking a local task and completing it in an efficient way,” Itamar says. Qodo’s plugin gives Codex access to codebase knowledge, organizational rules, and independent review through Qodo’s Agentic Toolbox.
With context, Codex can account for dependencies across repositories, including those hosted by different Git providers or services connected through Kafka. “Qodo will make that connection and will help Codex look at the software as a whole and not miss those P0 bugs,” Itamar says.
Before Codex starts writing code, it can use Qodo’s plugin to retrieve the relevant codebase rules and pull request history and patterns. After Codex makes the changes, Qodo reviews them and returns its findings directly to Codex in the same session, before the developer opens a pull request.
Nnenna Ndukwe, Qodo’s AI Developer Relations Lead, sees developers using this workflow to build features and tackle technical debt with GPT‑6 models in Codex. They are automating the handoff between Codex’s coding work and Qodo’s independent review. “That’s what I’ve been seeing developers do, which is really exciting because GPT‑6 is extremely capable,” Nnenna says.
Itamar describes coding and review agents as “yin and yang”: one creates code, the other challenges it. Each role needs its own setup and context and may involve multiple agents. His vision is for agents to discuss findings and “surface only what they don’t accept” for human review. Developers would become planners and reviewers, able to inspect unresolved issues alongside the agents’ discussion, context, and prior developer decisions.
“That’s how you create trust,” Itamar says. “I think it makes the whole process overall more efficient,” Nnenna adds. “We get to leverage the best parts of our expertise.”
Internally at Qodo, Nnenna has been using GPT‑6 Astra’s Extra High and Ultra effort settings in her own automated development workflows as well as knowledge work.
“What does Codex not do? I love everything that Codex does. I love the idea of combining AI and automation.”
Nnenna schedules Codex to report on what she accomplished during the week and where she could focus or improve next. She also uses it to research conversations about agentic engineering on social platforms and suggest ideas for developer content. When it comes to her own code changes, she schedules Codex to analyze her work and suggest agent skills to create or refine.
Itamar sees an opportunity to use OpenAI’s recently launched Decisions API to choose which rules and agent skills apply at each stage of development, rather than rely on language models for every decision.
He cites an example where the Decisions API could select which of 200 rules need checking before a language model reasons through the relevant ones, helping to make selection faster and cheaper while improving the quality of the review. “The Decisions API is an immense opportunity for us to bring more trust to the process,” Itamar says.
Qodo’s approach combines “the best aspects of artificial intelligence and the best aspects of traditional software engineering and scripting,” Nnenna says. Company knowledge guides coding agents, independent review checks their work, and human expertise resolves the questions that remain.
This creates the foundation of what Itamar describes as a software factory: a system customers could assemble and own, bringing coding and review agents together with an understanding of how software changes in production.
With OpenAI, Qodo is building a future where coding agents put more of a company’s experience to work and developers can focus their judgment where it matters most.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.