Tag
All content about Agent File, organized for fast scanning.
Recent developments in AI agent technology highlight advancements in memory management and task execution. New persistent memory features for subagents allow for better session continuity, while the introduction of a command for long-running tasks enhances the efficiency of autonomous workflows. Additionally, a significant overhaul of monorepo structures is reported to improve development speed and resource utilization for coding agents.
Lydia Hallie says subagents don’t inherit a main session’s auto memory, and need a "memory" field to persist across runs. She breaks down user, project, and local scopes, plus a workaround for mixing user and project memory.
xAI appears to be adding a new /goal command to Grok Build, aimed at long-running autonomous tasks using rounds of subagents to implement and verify work. The move echoes recent goal-based agent workflows in OpenAI Codex and similar tools.
Basis shared how it reorganized its monorepo to better support coding agents, built around “canonical” context and a six-layer AGENTS.md and skills architecture. The company claims the changes drove 5x higher token usage per developer and 2.5x faster weekly commits.