vercel/eve eve@0.79.0 on GitHub

bc2be45: Adds t.score(evaluation) for recording a score from 0 to 1, named with .label(key) and gated with .gate() or .atLeast(). AssertionResult gains key, and score, threshold, and passed are now optional: score is absent when the scorer threw, and threshold and passed are absent when no rule applies, so custom reporters should treat passed === false as the failure signal. eve eval warns when soft assertions in one eval share a name with different scores.

76f6854: eve's agent spans now come from the shared tracing library, and durable turns recover from one checkpoint per turn. Each tool call is still one execute_tool span, a local subagent's first turn still nests under its caller, and trace schema version 4 is unchanged.
f5e5183: Connection approval policies now receive toolAnnotations, the hints an MCP server declared for the called tool (such as readOnlyHint and destructiveHint), so a policy can skip approval for read-only tools. It is undefined when the server declares none.
302a4d3: Tools can now define approval.prompt to show a human-readable, input-aware summary in approval requests. Slack keeps the complete tool input available in a collapsed block below the approval card.
422b628: Extract reusable agent tracing with wrapped execution, model stream completion, remote delegation, and durable turns that save and resume themselves through a checkpointer. Preserve trace schema version 4 while sharing one span lifecycle with eve.
485d0f5: The agent tracing library now writes schema 4's agent.action.kind, agent.action.name, and agent.action.outcome on tool and approval spans itself, so hosts other than eve render correctly in Agent Runs. eve's trace output is unchanged.