Grok Build 1.0.14 hardens agent infrastructure

This is a meaningful maturity release: Grok Build is improving the operational plumbing that makes autonomous coding dependable, inspectable, and easier to integrate.

• –PostToolUse feedback lets validators, linters, and test runners return context directly to the model for self-correction
• –Persistent per-turn token and cost data makes agent spending easier to audit and optimize
• –Stronger subagent spawning, transcript resuming, and failure visibility reduce silent breakdowns in parallel workflows
• –Proactive OIDC refresh and faster startup target long-running sessions and large repositories
• –Per-effort model identifiers give teams more control over routing fast versus deep reasoning modes