Core FunctionalityCore Concepts
Reviews
Understanding code review output
Reviews
What a Review Contains
A Burg AI review includes:
- Inline comments on specific lines of code
- Issue description explaining the problem
- Severity level indicating urgency
- Suggested fix when applicable
Severity Levels
| Severity | Meaning | Examples |
|---|---|---|
| 🔴 Critical | Must fix before merge | Security vulnerabilities, data loss risks, crashes |
| 🟠 Major | Should fix | Bugs, performance issues, broken logic |
| 🟡 Minor | Consider fixing | Style issues, minor improvements, documentation |
What Is Considered In-Scope
The agent reviews:
- Logic errors and bugs
- Security vulnerabilities
- Performance anti-patterns
- Error handling gaps
- Type safety issues
- Code maintainability
What Is Ignored
The agent skips:
- Formatting (handled by linters)
- Naming preferences (subjective)
- Already-reviewed unchanged code
- Files matching ignore patterns
- Binary files
- Lock files and generated code
Review Scope Limits
To protect your API usage:
| Limit | Free | Pro |
|---|---|---|
| Max files per review | 25 | Configurable |
| Max file size | 300 KB | Configurable |
| Files auto-skipped | Binaries, locks, build output | Same |
If a PR exceeds limits, a subset of files is reviewed with a note explaining what was skipped.