Common Mistakes to Avoid
Learn from the community's experiences about what not to do when developing software with Cursor.
Sort By
Impact Level
Project Scale
Languages
Tools
Avoid Large Feature Drops
Don't request or implement large feature sets at once
Increases complexity and risk of errors
Feature implementation
Don't Skip Overview Phase
Avoid letting AI generate code immediately without understanding its planned approach
Can lead to wasted time debugging incorrect implementations and fixing misunderstandings
When starting new feature implementation with AI
Don't Assume AI Understanding
Avoid assuming AI has correctly understood all requirements without verification
Can lead to completely incorrect implementations despite fast code generation
During initial feature request to AI
Don't Rely Solely on AI Generated Code
Avoid blindly accepting AI suggestions without review
AI-generated code needs human verification for quality and security
When using Cursor AI for iOS development
Avoid Unguided AI Code Generation
Don't let AI generate code without clear test specifications or design guidelines
Can lead to unmaintainable or incorrect code that doesn't meet requirements
When using AI coding assistants for development
Avoid Direct Code Writing Without Tests
Don't write AI-generated code without test coverage in large projects
High risk of breaking existing functionality
When implementing features in established codebases
Don't Rely Solely on AI Tools
Avoid complete dependence on AI coding assistants for development tasks
Over-reliance on AI tools can lead to decreased programming skills and poor code quality
Software development workflow
Manual Context Repetition
Avoid repeatedly explaining project structure and context to AI tools
Wastes time, prone to inconsistencies, and reduces productivity
When working with AI assistants on code modifications
Uncontrolled AI Modifications
Prevent AI from making changes without understanding full project context
Leads to broken code and architectural inconsistencies
When using AI for code modifications
Avoid Single Model Dependency
Don't base your entire development strategy on a single AI model
Over-reliance on one model can limit capabilities and create single points of failure
When implementing AI-assisted development workflows
Avoid Undefined Design Requirements
Don't start development without clear design specifications
Leads to constant changes and extended development time
Initial project planning phase
Prevent Inconsistent Architecture Suggestions
Don't allow AI tools to suggest changes that break existing architecture
Can lead to architectural inconsistencies and technical debt
When receiving AI-generated code suggestions
Skip Changelog Updates
Don't commit changes without updating the changelog
Makes it difficult to track changes and can lead to incomplete release notes
Development workflow
Avoid Long Sessions Without Reset
Don't maintain extended AI sessions without periodic resets
Context degradation over time can lead to decreased AI performance and accuracy
During lengthy development sessions
Avoid Continuous Problematic Prompt Iteration
Don't spend excessive time iterating on prompts that aren't working
Wastes time and resources, often leads to frustration without improvement
When facing persistent issues with LLM outputs
Avoid Immediate Code Fixes
Don't jump directly into implementing code changes without proper analysis
Can lead to addressing symptoms rather than root causes, potentially creating more problems
When encountering a bug or issue
Avoid Using Cursor Without Version Control
Don't make AI-assisted changes without having a version control system in place
Lack of version control can lead to irreversible changes and code loss
When starting to use Cursor or any AI coding assistant
Avoid Vague Prompts
Don't use generic or unclear prompts like 'code stuff' with Cursor AI
Leads to hallucinated, random, and nonsensical code across various files
When initiating code generation with Cursor AI
Avoid Quick Conclusions
Don't rush to conclusions without thorough exploration
Premature conclusions can lead to overlooked issues and incomplete solutions
During problem-solving and analysis tasks
Avoid Static Schema References
Don't rely on static code snippets for database schema information
Leads to outdated information and incorrect code generation
When generating database-connected components