attac-t Skill: Package Positioning
"Build it because it's better, not because you can."
The Decision
Build when:
- No existing solution covers your use case
- Existing solutions have poor DX, are abandoned, or are unmaintained
- You can offer meaningfully better developer experience (not marginally -- meaningfully)
- The existing package's architecture prevents the improvement you need (contributing upstream won't work)
- You've identified a genuine gap, not just a preference
Contribute when:
- A well-maintained package exists with good DX
- Your improvement fits within the existing architecture
- The maintainer is responsive to PRs
- Your "new package" would be 80% identical to what exists
Wait when:
- You haven't installed and used the alternatives
- Your differentiation is theoretical, not tested
- The problem might not be worth solving as a package (see decide-extraction)
The Heuristic
Ask: "Can I explain in one sentence why a developer should choose this over what already exists?"
If no, either sharpen your differentiation or contribute to an existing package instead.
The Quick Test
| Ask | Answer | Action |
|---|---|---|
| Have you searched Packagist for alternatives? | No | Stop. Research first |
| Have you installed and used the top alternatives? | No | Stop. Evaluate DX first |
| Can you name your differentiation in one sentence? | No | Not ready to build |
| Is the best alternative abandoned (12+ months)? | Yes | Strong signal to build |
| Would your improvement fit as a PR upstream? | Yes | Contribute, don't compete |
| Does your positioning statement feel forced? | Yes | You don't have a real differentiator |
| Can you fill the DX comparison table honestly? | Yes | You've done the research |
Real-World Examples
See examples.md.