OKF v0.2：为智能体知识加上可信层

The Open Knowledge format has been improved. Google just announced OKF v0.2. which introduces a Trust Layer to solve the problem of accountability in agent-generated knowledge. OKF v0.2 aims to answer these 5 questions in the YAML frontmatter: ▪What was this created from? ▪How much should I trust it? ▪Is it still true? ▪Is it the current version? ▪Was this number produced the way we said it must be? Type is still the only required filed in the YAML frontmatter, but you are encouraged to add these new fields. (You don't need to though - your current OKF structure is still going to work.) The new fields are: ▪generated: (Supersedes timestamp) ▪verified: (List of human/machine confirmations) ▪status: (draft, stable, deprecated) ▪stale_after: (Absolute expiration date) ▪sources: (Structured provenance data) I want to share something else as well. I think if you're following how I'm building with OKF, we may be using OKF in a different way than it was intended to be used for. They say they created OKF so agents can have the context they need to understand table schemas, metric definitions and runbooks. Yet, we are using it to help agents understand knowledge. I think what we are doing is more akin to Karpathy's LLM-Wiki (which was the inspiration for OKF). I also think what we are doing is amazing - so continue on! https://t.co/9vrydTG1HS Marie