Teamharness Organization
Productivityteamharness-organizationTeamharness Organization is an Agent Skill for productivity published in the agentscope-ai/AgentTeams repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Use to understand TeamHarness team roles, responsibility boundaries, and who owns project, task, runtime, and control-plane work.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/agentscope-ai/AgentTeams/blob/HEAD/plugins/teamharness/skills/team/organization/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/teamharness-organization, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Organization
Use this skill to reason about team roles and responsibility boundaries.
Leader owns planning, delegation, acceptance, and requester reporting. Workers own assigned task execution. Remote members participate as invited local agents or humans. Manager owns control-plane operations.
Do not let a Worker silently become project owner. Do not let a remote member act like a controller-managed Worker unless the team explicitly models that role.
Role Boundaries
Leader owns:
- deciding the work breakdown
- creating project plans
- delegating task specs
- checking Worker results
- accepting or rejecting results for project progress
- reporting accepted outcomes to the requester
Worker owns:
- acknowledging assigned tasks
- reading
shared/tasks/{task-id}/spec.md - producing deliverables under
shared/tasks/{task-id}/ - submitting
result.md - reporting blockers or completion to the Leader
Remote member owns only the work explicitly assigned to that member. Treat remote members as participants, not controller-managed runtime resources.
Manager owns control-plane operations such as creating workers, changing model configuration, deploying teams, and managing runtime state. TeamHarness skills must not instruct agents to call control-plane APIs.
Safety Boundary
Never write secrets, credentials, tokens, authorization headers, or live runtime
state into TEAMS.md, project files, task specs, task results, or Matrix
messages. If a task needs credentialed access, refer only to the approved
credential name or environment variable name, not its value.