Skill v1.0.0

currentLLM-judged scan90/100

muqsitnawaz/.agents-system/teams

──Details

PublishedOctober 22, 2026 at 06:35 PM

Content Hashsha256:fc4889da98ef0e70...

Git SHA72ef06fd5882

──Files

Files (1 file, 3.1 KB)

SKILL.md3.1 KBactive

SKILL.md · 101 lines · 3.1 KB

version: "1.0.0" name: teams description: "Organize AI coding agents into teams that collaborate on a shared task. Create teams, add teammates, start them, monitor progress, and collect results. Use this skill when you need parallel agent execution. For single-agent dispatch, use agents run instead." argument-hint: "[create|add|start|status|disband]" allowed-tools: Bash(agents teams), Bash(agents run) user-invocable: true

Teams Skill

Organize AI coding agents into teams for parallel collaboration. This skill teaches you how to use the agents teams CLI.

Single Agent vs Teams

Single agent: Use agents run <agent> "prompt" --mode edit for one-off tasks
Multiple agents: Use agents teams when you need parallel execution

Quick Start

bash

# Create a team
agents teams create my-feature
 
# Add teammates
agents teams add my-feature claude "Implement the auth middleware" --name auth
agents teams add my-feature codex "Build the login UI" --name frontend
 
# Start the team
agents teams start my-feature --watch

Commands

Command	Description	Example
`create`	Start a new team	`agents teams create my-team`
`add`	Add a teammate	`agents teams add my-team claude "Task" --name role`
`start`	Launch pending teammates	`agents teams start my-team --watch`
`status`	Check who's working	`agents teams status my-team`
`logs`	Read teammate output	`agents teams logs my-team frontend`
`remove`	Remove a teammate	`agents teams remove my-team frontend`
`disband`	Stop all and remove	`agents teams disband my-team`
`doctor`	Check installed agents	`agents teams doctor`

DAG Dependencies

Use --after to create dependencies:

bash

# Backend first
agents teams add my-feature claude "Build API" --name backend
 
# Frontend waits for backend
agents teams add my-feature codex "Build UI" --name frontend --after backend
 
# QA waits for both
agents teams add my-feature claude "Run tests" --name qa --after backend,frontend
 
# Start drains the DAG automatically
agents teams start my-feature --watch

Modes

Mode	Use When
`plan` (default)	Read-only work: research, audit, analysis
`edit`	Code changes: implementation, refactoring

Always use --mode plan for security audits, research, and analysis.

Monitoring

bash

# Check status
agents teams status my-feature
 
# Delta poll (efficient)
agents teams status my-feature --since 2026-04-24T09:00:00-07:00
 
# Read one teammate's log
agents teams logs my-feature frontend

Best Practices

Mix agents if available — different agents have different blind spots
Use `--mode plan` for read-only work (audits, research)
Give full context — each teammate needs the big picture plus their specific task
Demand evidence — end prompts with: Return file:line quotes for every claim
Run in parallel — most tasks don't depend on each other
Name teammates with --name for easy reference

Short Aliases

teams c  = create    teams a  = add       teams s  = status
teams rm = remove    teams d  = disband   teams ls = list

All versions