[FEATURE] new cc agent-team！

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

Claude Code's native Agent mode is subject to three limitations: Agents cannot communicate directly with one another—only a parent agent can interact with its children, while sibling agents remain invisible to one another; it is restricted to a single LLM provider—meaning all agents must utilize the exact same API; and concurrent editing by multiple agents leads to conflicts—due to the absence of a file-locking mechanism, issues arise when two agents attempt to modify the same file simultaneously. Agent Team resolves these three issues, enabling you to assemble an AI team that is multi-model, capable of inter-agent communication, and equipped with conflict protection.
See here:https://github.com/hiddenguy1/cc-agent-teams

Proposed Solution

• Multi-Model Integration — Assigns different models to different roles (using more expensive models for decision-making and cheaper ones for execution).
• Direct Agent Communication — Enables any agent to send messages to any other agent, moving beyond the limitations of unidirectional leader-worker communication.
• File Conflict Protection — Implements atomic-level file locking to prevent multiple agents from simultaneously editing the same file.
• Automatic Wake-up — Automatically notifies the recipient immediately after a message is sent, eliminating the need for polling.
• Read Status Tracking — Each message features a status indicator to track whether it has been read or remains unread.

Alternative Solutions

• Multi-Model Integration — Assigns different models to different roles (using more expensive models for decision-making and cheaper ones for execution).
• Direct Agent Communication — Enables any agent to send messages to any other agent, moving beyond the limitations of unidirectional leader-worker communication.
• File Conflict Protection — Implements atomic-level file locking to prevent multiple agents from simultaneously editing the same file.
• Automatic Wake-up — Automatically notifies the recipient immediately after a message is sent, eliminating the need for polling.
• Read Status Tracking — Each message features a status indicator to track whether it has been read or remains unread.

Priority

Medium - Would be very helpful

Feature Category

Other

Use Case Example

_No response_

Additional Context

_No response_