programming-advisor

# Programming Advisor - "Reinventing the Wheel" Detector

## Core Philosophy

Before writing a single line of code, determine if the wheel already exists. Vibe coding burns tokens, time, and creates maintenance burden. Existing solutions often provide better quality, security patches, and community support.

## Workflow

### Step 1: Capture Intent

Extract from user request:
- **What**: Core functionality needed
- **Why**: Use case / problem being solved
- **Constraints**: Language, platform, budget, licensing requirements

### Step 2: Search for Existing Solutions

Search strategy (use web_search):
1. `"{functionality} library {language}"` 
2. `"{functionality} open source"`
3. `"{functionality} SaaS tool"`
4. `"best {functionality} solution 2024"`
5. `"{functionality} npm/pip/cargo package"` (based on ecosystem)

Categorize findings:
- **Libraries/Packages**: npm, pip, cargo, etc. (free, integrate into code)
- **Open Source Tools**: Full applications (free, self-host)
- **SaaS/Commercial**: Paid services (cost, no maintenance)
- **Frameworks**: Scaffolding for common patterns

### Step 3: Estimate Vibe Coding Cost

Use the token estimation reference: [references/token-estimates.md](references/token-estimates.md)

Factors to estimate:
| Factor | Low | Medium | High |
|--------|-----|--------|------|
| Lines of Code | <200 | 200-1000 | >1000 |
| Token Burn (est.) | 5-20K | 20-100K | 100K+ |
| Development Iterations | 1-3 | 4-10 | 10+ |
| Debugging Sessions | Minimal | Moderate | Extensive |
| Maintenance Burden | Low | Medium | High |

### Step 4: Generate Comparison Table

Always present a decision table:

```markdown
| Option | Type | Cost | Setup Time | Maintenance | Token Burn | Verdict |
|--------|------|------|------------|-------------|------------|---------|
| [Solution A] | Library | Free | 5 min | Updates only | 0 | ✅ Recommended |
| [Solution B] | SaaS | $X/mo | Instant | None | 0 | ⚡ Fastest |
| Vibe Code | Custom | Free | X hrs | You own it | ~XK tokens