agent-collaboration

Coordinate squad agents with worktree awareness and decision recording.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/CarlosSardo/nats-poc-simulator --skill agent-collaboration-carlossardo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-collaboration
Source: https://github.com/CarlosSardo/nats-poc-simulator/tree/main/.copilot/skills/agent-collaboration
Command: npx skills add https://github.com/CarlosSardo/nats-poc-simulator --skill agent-collaboration-carlossardo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Standardizes collaboration patterns for squad agents to enable predictable worktree awareness, decision recording, and cross-agent communication, reducing miscommunication and delays.

Core Features & Use Cases

  • Worktree Awareness: Use the TEAM ROOT path provided by your spawn prompt. All .squad/ paths are relative to this root. If TEAM ROOT is not provided (rare), run git rev-parse --show-toplevel as fallback. Never assume CWD is the repo root.
  • Decision Recording: After making a decision that affects other team members, write it to .squad/decisions/inbox/{your-name}-{brief-slug}.md
  • Cross-Agent Communication: If you need input, say so in your response; the coordinator will bring them in.
  • Reviewer Protocol: If you have reviewer authority and reject work: the original author is locked out from revising that artifact. A different agent must own the revision. State who should revise in your rejection response.
  • Anti-Patterns: Do not read all agent charters; do not write directly to .squad/decisions/inbox; do not modify other agents' history.md; never assume CWD is repo root.

Quick Start

Record a new collaboration decision in .squad/decisions/inbox using the prescribed header and content format.

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 multi-agent squad workflow?▼

Standardize cross-agent communication by routing input requests through a coordinator and recording decisions as markdown artifacts in a shared inbox. This ensures squad agents share input and coordinate responses predictably without direct messaging.

How does worktree awareness work when coordinating squad agents?▼

Worktree awareness works by using the TEAM ROOT path from the spawn prompt, keeping all .squad/ paths relative to this root. If TEAM ROOT is missing, agents run git rev-parse --show-toplevel as a fallback to avoid assuming the CWD is the repo root.

What is the best way to record decisions that affect other team members?▼

Record decisions by writing a markdown file to .squad/decisions/inbox/{your-name}-{brief-slug}.md using the prescribed header and content format. This standardizes authoring decisions so other squad agents can discover and react to them.

What happens when a reviewer rejects work in a squad workflow?▼

When a reviewer rejects work, the original author is locked out from revising that artifact. A different agent must own the revision, and the reviewer must state who should revise it in their rejection response.

What are the anti-patterns to avoid when managing multi-agent collaboration?▼

Avoid reading all agent charters, writing directly to .squad/decisions/inbox without naming conventions, modifying other agents' history.md files, and assuming CWD is the repo root. These anti-patterns cause miscommunication and delays.

Do I need external runtime dependencies to enforce squad agent collaboration patterns?▼

No external runtime dependencies are required. The collaboration patterns are enforced entirely through inbox-based decision artifacts and TEAM ROOT conventions defined in SKILL.md, requiring no additional software installations.