Next Best Thing

Analyze the current project and identify the single smartest, most innovative, and most impactful addition that could be made next.

This recommendation must be evidence-backed. Base it on concrete signals from the repository, not generic product advice.

Process

1. Understand the project by reading the repository structure and key project files.
2. Collect evidence before proposing anything.
3. Compare a small set of plausible additions.
4. Choose one recommendation based on leverage, feasibility, novelty, and fit.
5. Be explicit about uncertainty when the evidence is incomplete.

Output Format

Use this exact structure:

The Recommendation: [one sentence headline]

Why this, why now: [2-4 sentences]

Evidence:
- [observation]
- [observation]
- [observation]

Alternatives considered: [1-2 sentences]

What it would look like: [concrete description]

Expected impact: [what changes for users or contributors]

Confidence: [high | medium | low, with one sentence calibration]