The latest articles on DEV Community by Hamza KONTE (@hamza_konte_04e855f6449da). https://dev.to/hamza_konte_04e855f6449da https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3823414%2F4c8f3c47-b138-4872-b260-d2caa4e7b897.jpeg https://dev.to/hamza_konte_04e855f6449da en Hamza KONTE Sat, 14 Mar 2026 05:53:42 +0000 https://dev.to/hamza_konte_04e855f6449da/a-colleague-told-me-my-prompt-was-garbage-he-was-right-59j4 https://dev.to/hamza_konte_04e855f6449da/a-colleague-told-me-my-prompt-was-garbage-he-was-right-59j4 <p>Role, constraints, examples, output format. All in one paragraph. The model has to guess what's what.</p> <p>I kept rewriting my prompts, re-running, tweaking one sentence, re-running again. Ten tries to get something decent. Not because the model was bad. Because my instructions were a mess.</p> <h2> The problem nobody talks about </h2> <p>When you write a prompt, you think in categories. You know which part is the role, which part is a constraint, which part is an example. But the model doesn't. It sees one block of text and has to infer the boundaries.</p> <p>Sometimes it gets it right. Sometimes it treats your example as an instruction. Sometimes it ignores your output format because it blended into the context paragraph above it.</p> <p>The fix isn't a better model. It's showing the model where each piece starts and ends.</p> <h2> So I built flompt </h2> <p><a href="https://flompt.dev" rel="noopener noreferrer">flompt</a> pre-interprets your prompt before the model does.</p> <p>You paste your prompt. AI splits it into typed blocks: role, audience, objective, constraints, examples, chain of thought, output format, response style, language. You see exactly how your instructions will be read.</p> <p>If the interpretation is wrong, you fix the blocks. Then you compile and get a structured prompt that the model parses correctly on the first try.</p> <p><strong>Same model. Same task. Night and day.</strong></p> <h2> What it looks like </h2> <p> <iframe src="https://www.youtube.com/embed/hFVTnnw9wIU"> </iframe> </p> <h2> Three ways to use it </h2> <p><strong>Web app</strong> at <a href="https://flompt.dev" rel="noopener noreferrer">flompt.dev</a>. No account, no install. Paste, decompose, edit, compile.</p> <p><strong>Browser extension</strong> for Chrome and Firefox. Adds an "Enhance" button right inside ChatGPT, Claude, and Gemini. You structure your prompt in a sidebar without leaving the conversation.</p> <p><strong>MCP server</strong> for Claude Code. Your agents can call <code>decompose_prompt()</code> and <code>compile_prompt()</code> programmatically:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>claude mcp add <span class="nt">--transport</span> http <span class="nt">--scope</span> user flompt https://flompt.dev/mcp/ </code></pre> </div> <h2> 50+ templates </h2> <p>Don't want to start from scratch? There are 50+ templates built from the best prompts on GitHub. Code review, blog post, sales pitch, data analysis, lesson plan. Load one, tweak the blocks, compile.</p> <h2> Stack </h2> <p>React 18 + TypeScript + React Flow + Zustand (frontend), FastAPI + Python 3.12 (backend), Caddy (reverse proxy). 10 languages supported.</p> <p>Free. Open-source. MIT.</p> <p>Try it: <a href="https://flompt.dev" rel="noopener noreferrer">flompt.dev</a><br> Repo: <a href="https://github.com/Nyrok/flompt" rel="noopener noreferrer">https://github.com/Nyrok/flompt</a></p> <p>If you find it useful, a star on the repo would mean a lot. Solo project, every star helps with visibility.</p> ai promptengineering opensource productivity