Evidence observed in google/adk-go, a Agent Infrastructure project.
Related friction appears in 3 independent repositories. This is stronger than one backlog item, but still requires direct user validation.
Project strength and demand confidence are measured separately.Reporter context: When will skills be supported? This is very important.Excerpted from the public Issue. Read the complete thread before interpreting it.
Recruit: Recruit teams with live traces, tool calls or deployment constraints from a real agent workflow.
Guardrail: Test observable state transitions, permission boundaries and recovery from partial failure.
For [builder], the agent cannot reliably [select, add, call or remove a tool] during [workflow], causing [manual intervention or failed work].
You can name one user, one situation and one measurable consequence without proposing a feature.Implement or simulate one tool lifecycle operation with explicit inputs, outputs and failure handling, then replay three real calls.
Three representative calls complete with the expected state transition and an observable recovery path.Test observable state transitions, permission boundaries and recovery from partial failure.Information has value only when it changes action. This page turns one public signal into a bounded validation exercise. It is a research aid, not proof of demand, investment advice or a product recommendation.