01
Overall cleanliness score
A simple overall score helps developers quickly judge whether a prompt looks clean and ready or still needs revision.
VS Code extension for prompt quality analysis
Analyze prompt quality with a radar chart and clear feedback.
Prompt Radar helps developers inspect prompts embedded in source code and score them across five categories: Security, Safety, Formatting, Efficiency, and Reliability. Higher scores mean the prompt looks cleaner and stronger in that dimension.
Explore the GitHub repo Explore the source code on GitHub.3.2
What this result means
The prompt is efficient and reasonably well formatted, but it still needs stronger safety and security guidance. Prompt Radar combines the five category scores into one overall cleanliness score.
Formatting: Strong
Efficiency: Strong
Security: Needs work
About the tool
Prompt feedback directly where developers write prompts.
Prompt Radar is designed for prompts-as-code. A developer selects a prompt in the editor, runs the analysis, and gets a score-driven view that highlights strengths, weaknesses, and actionable improvement suggestions inside VS Code.
02
Five category scores
The radar chart shows how the prompt performs across Security, Safety, Formatting, Efficiency, and Reliability.
03
Detailed feedback and assessment
Each category comes with an explanation and a lightweight user assessment so the tool can be reviewed and improved.
Scoring categories
Higher values in the radar mean better prompt quality.
Prompt Radar scores each dimension from weak to strong. A higher score means the prompt appears cleaner, more explicit, safer, and more robust in that category.
FormattingClear structure, explicit output expectations, and parseable responses.
ReliabilitySpecific, stable instructions that reduce ambiguity and inconsistent behavior.
EfficiencyConcise wording and useful context without unnecessary token cost.
SecurityReduced risk of leakage, unsafe access, injection, or sensitive exposure.
SafetyLower risk of harmful, unethical, biased, or policy-problematic outputs.
4.3
Clean prompt
A well-written prompt with clear formatting, precise instructions, and balanced coverage across all categories.
3.1
Needs refinement
Good structure and moderate efficiency, but weaker reliability and security signals reduce the overall score.
2.4
Revision recommended
The prompt may be acceptable in one area, but it still needs clearer structure and stronger protections before use.
Workflow
Analyze, review, and refine prompts inside VS Code.
Select a prompt
The developer highlights a prompt, template, or message directly in the editor.
Run Prompt Radar
The extension analyzes the selection using the LLM-based scoring pipeline.
Review the score
The overall score and radar chart summarize how strong the prompt looks across the five categories.
Use the feedback
The developer reads the detailed analysis, gives quick assessment feedback, and improves the prompt.
Analysis view
A simple detailed analysis page for each prompt.
After analyzing a prompt, the extension can show the overall score, the radar chart, a short explanation, and a category-by-category assessment panel that lets users agree, disagree, or stay unsure.
3.4
An embedded prompt for summarization. It still needs stronger protection against untrusted input and more precise security constraints, but its formatting and safety signals are relatively strong.
Formatting - 3.5 / 5
ModerateThe prompt defines the task clearly but could be more explicit about the final response structure and length.
Provide a concrete output format, such as bullet points, a short paragraph, or strict JSON.
Security - 1.7 / 5
Needs workThe prompt does not sufficiently constrain how untrusted or injected content should be handled.
Add instructions that ignore malicious embedded directions and avoid revealing secrets or internal context.
Missed smells
Developer feedbackIf Prompt Radar misses an issue, the developer can add a missing smell with its category, severity, short name, and an optional explanation.
Formatting
Minor
Demo video
See the full prompt analysis workflow.
Watch prompt selection, extension execution, the radar chart, the detailed analysis, and the user assessment flow.
Developer feedback
Help us improve Prompt Radar.
The extension includes per-category assessment, optional rationale, and a way to add missed smells that were not detected. There is also a survey for developers who use the extension to assess Prompt Radar's installation experience, ease of use, reliability, and overall feedback.