Contributing

Thanks for taking the project seriously.

The main rule is: do not make weak sources look stronger than they are.

Development Setup

git clone https://github.com/qiaozhi7426-gif/source-to-skill.git
cd source-to-skill
python -m venv .venv
. .venv/bin/activate
python -m pip install -e . pytest
python -m pytest -q

What Makes A Good Contribution

Good changes usually improve one of these:

• readiness scoring accuracy
• clearer artifact boundaries
• evidence preservation
• source intake quality
• tests with realistic weak and strong examples
• documentation that reduces overclaiming

What To Avoid

• Do not claim every source can become a full skill.
• Do not hide scoring logic behind an opaque model call.
• Do not add a new source type unless it normalizes into text before scoring.
• Do not add glossy marketing copy that makes the project sound magical.
• Do not mix private source examples into the public repo.

Test Expectations

Add tests when changing scoring, output-level thresholds, or generated file layouts.

Useful tests compare at least two cases:

• a weak source that should stay Note or Seed
• a strong source that can become Mini or Full Skill

Copy Style

Keep prose direct and conservative. Prefer:

This source is not strong enough for a full skill.

Over:

Transform any knowledge into a powerful AI skill instantly.