The latest articles on Forem by Umar F Yunusa (@umar_fyunusa_ab97415e208). https://forem.com/umar_fyunusa_ab97415e208 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%2F3611029%2F5c32ab67-f9df-415f-801e-b56bbb3c101a.jpg https://forem.com/umar_fyunusa_ab97415e208 en Umar F Yunusa Fri, 14 Nov 2025 09:27:55 +0000 https://forem.com/umar_fyunusa_ab97415e208/i-built-a-cli-tool-that-analyzes-llm-prompts-for-clarity-cost-and-security-would-love-feedback-d91 https://forem.com/umar_fyunusa_ab97415e208/i-built-a-cli-tool-that-analyzes-llm-prompts-for-clarity-cost-and-security-would-love-feedback-d91 <p>I kept iterating on prompts without knowing if my changes were making them better or worse, so I built a tool to help.</p> <p>PromptLint is a CLI that analyzes prompts before you send them to LLMs. It gives you three scores:</p> <p>Clarity (0-10)</p> <ul> <li>Detects ambiguous phrases like "maybe", "try to", "as needed"</li> <li>Checks for clear instructions and output format</li> <li>Flags conflicting instructions.</li> </ul> <p>Cost Efficiency</p> <ul> <li>Counts tokens using tiktoken</li> <li>Estimates costs across 7 models (GPT-4o, Claude, Gemini, etc.)</li> <li>Shows you the $ impact before you run.</li> </ul> <p>Security</p> <ul> <li>Detects prompt injection patterns</li> <li>Flags unvalidated user input</li> <li>Identifies risky operations</li> </ul> <p><a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fqezp4szggl0d2nf022hb.png" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fqezp4szggl0d2nf022hb.png" alt=" " width="800" height="733"></a></p> <p>✓ This change improves the prompt It's free and open source.<br> <code>Install: pip install promptlint</code><br> GitHub: <a href="https://github.com/fyunusa/promptlint" rel="noopener noreferrer">https://github.com/fyunusa/promptlint</a><br> PyPI: <a href="https://pypi.org/project/promptlint%60" rel="noopener noreferrer">https://pypi.org/project/promptlint`</a></p> <p>Would love feedback on:</p> <ul> <li>What other analysis dimensions would be useful?</li> <li>Are the clarity heuristics accurate?</li> <li>What models should I add support for? Happy to answer questions!</li> </ul> llm tooling cli showdev