Cursor Search & Analysis Tool

This skill enforces a rigorous discovery-first methodology for code implementation in Cursor. It prevents code duplication and ensures architectural consistency by mandating comprehensive searches before any code changes.

Core Principles

1. **Search Before Code**: Never write or suggest code without first searching the codebase

2. **Document Everything**: All findings must be documented in a structured format

3. **User Approval Required**: Wait for explicit approval before implementing changes

4. **Prefer Enhancement**: Always enhance existing code over creating new files

5. **Justify New Code**: Creating new files requires detailed justification

Implementation Workflow

Phase 1: Discovery (MANDATORY)

Before ANY implementation work, execute these searches:

```bash

Search for existing functionality

grep_search <relevant_keywords>

Find similar files

file_search <related_names>

Explore directory structure

list_dir <relevant_directories>

```

Phase 2: Documentation (MANDATORY)

Create a comprehensive findings report:

```markdown

Existing Files Found

Existing Functionality

Gaps Identified

Implementation Decision Tree

Proposed Approach

Implementation Plan

1. [Detailed first step with effort estimate]

2. [Detailed second step with effort estimate]

3. [Additional steps as needed]

```

Phase 3: Approval (MANDATORY)

Phase 4: Implementation

Only after approval:

1. Follow the approved implementation plan exactly

2. Update documentation as you progress

3. Run tests and validations at each step

4. Document any deviations from the plan

Search Commands Reference

grep_search

Search for code patterns, function names, or text across the codebase:

```bash

grep_search "function_name"

grep_search "import.*Component"

grep_search "TODO|FIXME"

```

file_search

Find files by name or pattern:

```bash

file_search "component.tsx"

file_search "*.config.js"

file_search "test"

```

list_dir

Explore directory structure:

```bash

list_dir src/

list_dir components/

list_dir .

```

Anti-Patterns to Avoid

❌ **NEVER DO THESE**:

Examples

Example 1: Adding a New Feature

**User Request**: "Add authentication middleware"

**Correct Approach**:

```bash

Step 1: Search

grep_search "middleware"

grep_search "authentication"

file_search "*middleware*"

list_dir src/middleware/

Step 2: Document findings

- Found existing auth.middleware.ts in src/middleware/

- Has JWT validation but missing refresh token logic

- Uses consistent error handling pattern

Step 3: Propose

"I found existing authentication middleware at src/middleware/auth.middleware.ts.

It has JWT validation but is missing refresh token support.

Options:

1. Enhance existing middleware with refresh token logic (recommended)

2. Create separate refresh middleware (only if different auth flow needed)

Which approach would you prefer?"

Step 4: Wait for approval before coding

```

Example 2: Bug Fix

**User Request**: "Fix the CSV parsing bug"

**Correct Approach**:

```bash

Search for CSV-related code

grep_search "csv"

grep_search "parse"

file_search "*csv*"

Document all CSV parsing locations

Identify which file has the bug

Propose targeted fix

Get approval

Implement fix

```

Integration with GitHub Actions

When working with CI/CD pipelines:

1. Search for existing workflow files: `.github/workflows/`

2. Document current automation setup

3. Identify gaps in testing or deployment

4. Propose enhancements to existing workflows

5. Create new workflows only if justified

Best Practices

Success Criteria

✅ A successful implementation includes:

Constraints

---

**Remember**: The most important rule is to SEARCH FIRST. Code duplication is worse than spending extra time on discovery. When in doubt, search more and ask the user for guidance.