The Crystal Ball

Goal: Identify what the codebase COULD BE, not what it IS.

Step 1: Load Reference

Read references/roadmap-patterns.md for presentation patterns and anti-patterns.

Analysis Checklist

1. Complexity Analysis (Priority)

For key algorithms/functions:

• Time complexity: Count loops, recursion, data structure ops
• Space complexity: Memory usage patterns
• Best/worst case: Performance characteristics

Document in plain English: "This search is O(n²) because it compares every item to every other item. Could be O(n log n) with sorting first."

2. Performance Opportunities

Look for:

• Unnecessary iterations that could be reduced
• Missing caching for expensive operations
• Synchronous operations that could be parallel
• Memory allocations that could be pooled
• Network calls that could be batched

3. Code Quality Issues

Detect:

• Dead code (unused exports, commented blocks)
• Outdated dependencies (check package.json versions)
• Code smells (long functions >50 lines, deep nesting >3 levels)
• TODO/FIXME comments (extract and list)
• Inconsistent patterns (mixed async styles, naming conventions)

4. Security Hardening

Identify:

• Missing input validation on public endpoints/CLI args
• Hardcoded secrets (API keys, passwords in code)
• Unsafe dependencies (check for known vulnerabilities)
• Missing rate limiting on public endpoints
• SQL injection or XSS vulnerabilities

5. Technical Debt

Find:

• Deprecated APIs still in use
• Inconsistent patterns across codebase
• Missing tests for critical paths
• Documentation gaps (CONTRIBUTING.md, examples/)
• Workarounds marked with comments

6. Feature Gaps

Based on project type, identify:

• Common features for this type that are missing
• User-requested features (check GitHub issues if accessible)
• Competitive features (what similar tools have)

Output Format

REQUIRED: Return the analysis in this specific markdown format so pen-wielding can read it.

# Roadmap Candidates

## Performance Opportunities
| Location | Current | Opportunity | Impact |
|----------|---------|-------------|--------|
| `src/search.ts:45` | O(n²) nested loops | Use Map for O(1) lookup | High |
| `src/api.ts:120` | Sequential API calls | Batch with Promise.all | Medium |

## Technical Debt
| Location | Issue | Suggested Fix |
|----------|-------|---------------|
| `src/utils.ts:30` | TODO from 6 months ago | Implement or remove |
| `package.json` | lodash@3.x outdated | Upgrade to 4.x |

## Feature Gaps
| Feature | Why It Matters | Effort |
|---------|----------------|--------|
| Config file support | Users want persistent settings | Medium |
| Fluent data access | Stop two-step retrieval dance | Low |

## Security Hardening
| Location | Risk | Recommendation |
|----------|------|----------------|
| `src/cli.ts:15` | No input sanitisation | Validate before use |

## Complexity Notes
| Function | Complexity | Plain English |
|----------|------------|---------------|
| `findMatches()` | O(n²) | Compares every item - slow for large lists |

## Community Help Wanted
| Task | Context |
|------|---------|
| Add Tests | `src/core/` has 0% coverage |
| Examples | Create `examples/` folder with basic usage |

Important

• Focus on README-worthy improvements (things users/contributors care about)
• Be specific with file:line references
• Explain WHY something is an opportunity, not just WHAT
• Prioritise by impact (High/Medium/Low)
• Don't overwhelm - top 5-10 items per category max

Handoff

Goal: Transition from technical analysis to creative brainstorming.

1. Present Findings: Show the "Roadmap Candidates" tables.
2. Verify: Ask the user:

   "The Crystal Ball has spoken. Happy with these insights, or want to explore more? Ready for a brain-jam with Gemini?"

3. Transition:
   • If user wants changes: Refine the analysis.
   • If user is satisfied: "Proceeding to brain-jam."