what is lazyprompt fr?

lazyprompt is a plugin and skill for OpenAI's Codex coding assistant that turns a vague, one-line idea into a well-formed prompt for two other Codex skills, ulw-plan and ulw-loop, from a related project called lazycodex or OmO. It does not run any code or carry out the task itself, it only writes the prompt text for you to copy and paste. You type a command starting with lazyprompt followed by a plain description of what you want, such as adding a search feature to your app. The skill decides whether your idea is specific enough to jump straight to an execution-and-verification prompt, or vague enough that it should first produce a planning prompt. If it is inside a project folder, it first reads any existing agent instruction files, such as AGENTS.md, CLAUDE.md, or cursor rule files, and folds those rules into the prompt it writes rather than asking about things already documented there. When your idea is vague, lazyprompt can automatically switch into a short interview mode, asking one clarifying question at a time about the weakest part of your idea rather than guessing, and you can request a deeper interview or skip straight to output instead. The output always has two parts: the actual prompt meant to be copied into ulw-loop or ulw-plan, and a separate plain-language explanation of what that prompt does, which is meant only for you to read and not to paste anywhere. Depending on the task, it can also weave in optional helper skills, such as ones for frontend work, structural code search, or checking references and errors, only including a helper if it actually fits the task. The project explicitly avoids inventing success criteria, constraints, or helper skills that were not confirmed, and it never runs the target skill on your behalf. It is installed as a Codex plugin from its GitHub repository and is released under the MIT license.