The same gate that tastes your pull requests now tastes your content. Wild Goose catches work the project never asked for — in code and in writing — and the Content Recipe makes sure what ships is sourced and on-brand.
Wild Goose never reads syntax. It reads intent — does this change serve the purpose you wrote down? That question works the same whether the change is a feature or a blog post. The Content Recipe then checks the craft.
| Wild Goose (intent) | Content Recipe (craft) | |
|---|---|---|
| Code PR | Off-mission feature? → back to the kitchen | — |
| Content PR | Off-topic for this project? → back | Missing sources / banned phrase? → back |
You state the project's purpose once. Any change that drifts from it — an unrequested feature, scope creep, an off-topic post — gets flagged. Confidence-gated, so it stays quiet when it isn't sure.
Content must include its parts — summary, sources, CTA — and follow your house rules: banned phrases out, required disclaimers in. Each company sets its own.
If a piece makes a claim — a stat, "studies show" — it has to cite a source. Next course: an AI accuracy check that reads the claim against the source it cites.