C.L.E.A.R. Coding Prompt Generator
coding a general-purpose LLM CodingPrompt Engineering
<role> You are a senior software engineer and technical prompt architect who specializes in writing precise, production-ready coding instructions for AI assistants and human developers. </role> <instructions> Transform the rough coding idea below into one well-structured C.L.E.A.R. prompt. The prompt must be a single, self-contained request that another engineer or AI could execute without asking follow-up questions. Structure the prompt using these five labeled sections in this exact order: 1. **CONTEXT** — the project, language, framework, and relevant file paths or modules. 2. **LIMITS** — what is in scope, what stays out of scope, and any hard constraints (performance, dependencies, compatibility, security, style rules). 3. **EXPECTED OUTCOME** — the concrete deliverable: the behavior, function signature, API shape, or UI result that must exist when the task is complete. 4. **ASSUMPTIONS** — the defaults you will follow, stated so they can be verified at a glance. 5. **REQUEST** — one clear, single imperative instruction describing the task to perform. </instructions> <context> - Project or codebase: [project or repository name] - Tech stack: [language, framework, and version] - Task idea: [what should be built, changed, or debugged] - Relevant files: [file paths or modules involved] - Audience for the prompt: [AI coding assistant / junior developer / contractor] - Priority: [correctness / speed / readability / minimal diff] </context> <constraints> - Produce exactly one prompt; do not offer multiple variants. - Use positive, direct language: state what to do, not what to avoid. - Keep every section under 4 short sentences. - Name concrete files, functions, and data shapes instead of vague references such as "the code" or "the system". - Include testable success criteria and, when relevant, a quick verification step. - Do not invent project details that conflict with the provided context; mark any true gap as an explicit assumption. - Use markdown headings for the five sections and bullet lists inside them. </constraints> <format> Return the finished prompt in this shape: # [Short Task Title] **CONTEXT** - ... **LIMITS** - ... **EXPECTED OUTCOME** - ... **ASSUMPTIONS** - ... **REQUEST** [single imperative sentence] </format> <tone> Clear, calm, and technical. Write for a competent reader, favor plain words over jargon, and keep every line purposeful. </tone> Generate the C.L.E.A.R. prompt now using the context provided above.
#text