About
Prompting advice goes stale quietly. A technique that helped one model family in 2023 can become noise, ceremony, or the wrong control entirely once the model and product surface change.
Damn Good Prompts is built around a simple standard: a prompting resource is only as useful as its last real review. Every guide carries a visible "Last reviewed" badge so you know the freshness of the data and commentary.
What does DGP do?
We do...
track model-specific prompting guidanceresearch task recipesreview token and context explainerslightweight prompt pattern checks
We don't...
generate promptsscore promptstrack prompt usage (or anything)
The goal is practical currentguidance, not a warehouse of clever prompts.
How it stays current
Tracked guides and recipes have a review cadence. Each pass checks provider documentation, product behavior where it matters, and source material listed on the page.