agent-collaboration

Standardize cross-agent collaboration with worktree awareness and inbox-based decision recording.

3|3|Updated Apr 13, 2026
One-click install
npx skills add https://github.com/quaid-app/quaid --skill agent-collaboration-quaid-app
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-collaboration
Source: https://github.com/quaid-app/quaid/tree/main/.copilot/skills/agent-collaboration
Command: npx skills add https://github.com/quaid-app/quaid --skill agent-collaboration-quaid-app

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Centralizes collaboration practices across squad agents to ensure consistent worktree context, standardized decision capture, and reliable cross-agent communication, reducing friction and misalignment in multi-agent workflows.

Core Features & Use Cases

  • Worktree Awareness: Read TEAM ROOT from your spawn prompt and operate with paths relative to the repo root; never assume the current directory is the repository root.
  • Decision Recording: After making a decision affecting teammates, write it to .squad/decisions/inbox/{your-name}-{brief-slug}.md with a header, author, and rationale.
  • Cross-Agent Communication: If input from another agent is needed, ask for input and let the coordinator bring them in, preserving a centralized review flow.
  • Reviewer Protocol: If you have authority but disagree, assign a different owner for revisions and clearly state who should revise in your rejection.
  • Anti-Patterns: Avoid reading all agents' charters; avoid writing directly to decisions.md outside the inbox; avoid editing other agents' history.md; avoid assuming CWD is repo root.

Quick Start

Enable squad-wide collaboration by adopting the inbox-based decisions pattern and the worktree-aware spawn prompts.

Frequently Asked Questions about agent-collaboration

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I standardize cross-agent communication in a shared worktree?

Standardize cross-agent communication by applying worktree-aware spawn prompts, routing decisions to a shared .squad inbox, and using a coordinator to facilitate input. This ensures consistent behavior and predictable, auditable workflows across squad agents.

What is the best way to record decisions that affect other agents in a multi-agent workflow?

Record decisions by writing a markdown file to .squad/decisions/inbox/{your-name}-{brief-slug}.md, including a header, author, and rationale. This inbox-based pattern captures decisions consistently without directly editing the main decisions.md file.

How do I handle disagreements during a multi-agent code review?

Handle disagreements during multi-agent code review by assigning a different owner for revisions. Clearly state who should revise the work in your rejection to preserve a centralized review flow and maintain authority boundaries.

Why should agents not assume the current directory is the repository root?

Agents should not assume the current directory is the repository root because worktree contexts vary. Agents must read the TEAM ROOT from their spawn prompt and operate with paths relative to the repo root to avoid file path resolution errors.

What are the limitations of direct cross-agent collaboration without a coordinator?

Without a coordinator, agents risk reading all charters, editing other agents' history.md, or writing directly to decisions.md. These anti-patterns create friction and misalignment, bypassing the centralized review flow needed for auditable workflows.