skills.rest
agentscope-ai
agentscope-aiOFFICIAL
Explore all skills by this author →
𝕏@agentscope_aiGitHub

teamharness-organization

Defines team roles and responsibility boundaries for multi-agent collaboration workflows.

5.5k|677|Updated Feb 21, 2026
Search Tags:#delegation#Collaboration#control-plane#team roles#task ownership#responsibility boundaries#Multi-Agent
One-click install
npx skills add https://github.com/agentscope-ai/HiClaw --skill teamharness-organization
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: teamharness-organization
Source: https://github.com/agentscope-ai/HiClaw/tree/main/plugins/teamharness/skills/team/organization
Command: npx skills add https://github.com/agentscope-ai/HiClaw --skill teamharness-organization

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

In multi-agent teams, unclear ownership causes Workers to act as project owners, remote members to behave like managed runtime resources, and agents to overstep into control-plane operations. This Skill gives agents a clear model of who owns what so delegation, execution, and reporting stay consistent.

Core Features & Use Cases

  • Role Boundary Definitions: Clarifies that the Leader owns planning, delegation, acceptance, and requester reporting, while Workers own task execution and result submission.
  • Remote Member and Manager Scoping: Treats remote members as participants with explicitly assigned work, and reserves control-plane operations (creating workers, changing model configuration, deploying teams) for the Manager.
  • Safety Boundary Enforcement: Prohibits writing secrets, credentials, tokens, or live runtime state into TEAMS.md, task specs, results, or Matrix messages.
  • Use Case: When a Leader agent delegates a task, it uses this Skill to write the spec to shared/tasks/{task-id}/spec.md, let the Worker produce deliverables and result.md, and then accept or reject the outcome before reporting to the requester.

Quick Start

Use the teamharness-organization skill to explain which team role owns task delegation, result acceptance, and control-plane operations in this project.

Frequently Asked Questions about teamharness-organization

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

FAQPage Schema
How do I assign tasks between Leader and Worker agents?▼

The Leader creates the project plan and writes task specs to shared/tasks/{task-id}/spec.md, then delegates them. Workers acknowledge the task, produce deliverables under the same directory, submit result.md, and report blockers or completion back to the Leader.

What is the difference between a Worker and a remote member in a multi-agent team?▼

A Worker is a controller-managed runtime resource that executes delegated tasks, while a remote member is an invited local agent or human who only owns explicitly assigned work. Remote members should be treated as participants, not as managed runtime resources.

Can a Worker agent take over project ownership from the Leader?▼

No. The Leader owns planning, delegation, acceptance of results, and reporting to the requester. A Worker must not silently become the project owner; it only executes assigned tasks and reports outcomes.

Should agents call control-plane APIs to create workers or change models?▼

No. Control-plane operations such as creating workers, changing model configuration, deploying teams, and managing runtime state belong to the Manager. TeamHarness skills must not instruct agents to call control-plane APIs.

How should credentials be handled in task specs and team messages?▼

Never write secrets, tokens, authorization headers, or live runtime state into TEAMS.md, task specs, results, or Matrix messages. Refer only to the approved credential name or environment variable name, never its value.

Related Skills

Explore More Software Engineering Skills→
agentscope-aiagentscope-ai
5.3k

team-management

Create and manage agent teams with a Team Leader orchestrating workers.

Official
Intermediate
indigoai-usindigoai-us

execute-task

Execute a PRD story through a Ralph-pattern workflow with worker handoffs.

Official
Advanced
datamonsterrdatamonsterr

speckit-sync-conflicts

Detect contradictory requirements across specification and design documents.

Community
Advanced
dotBeepsdotBeeps

quest

Coordinate subagent work across ally roles with dispatch modes.

Community
Advanced