Wire flash
Claude Code launches Projects feature with Claude Tag-style agent architecture
Editorial responsibility
- No named human review is recorded for this page.
- Source reporting is collected, normalized, translated or condensed automatically when needed.
- Automatically published source-backed update
A new feature called Projects has been introduced for Claude Code, bringing the architecture of Claude Tag to the coding assistant. The feature assigns one agent per project that manages memory and spawns subagents for tasks. Users can configure the agent to be proactive and perform actions on a schedule. The post describes the experience as much nicer than juggling multiple sessions and encourages users to try it. This represents a significant enhancement to Claude Code's capabilities by introducing persistent, project-specific agent memory and task delegation.
Source report
The architecture of Claude Tag has been integrated into Claude Code.
Key Features
- One agent per project – manages memory and spawns subagents for tasks
- Proactive behavior – can be instructed to take initiative
- Scheduled actions – performs tasks on a schedule
This approach provides a smoother experience than managing multiple sessions. Try it out!
Source
trq212Neutral / independent
Part of this Story
Anthropic launches Claude Code Projects with parallel threads and persistent cloud sessions