coordinator-loop

Directs bounded implementer work through canonical Projectman assignments and reviews.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/eeemzs/aops --skill coordinator-loop-eeemzs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: coordinator-loop
Source: https://github.com/eeemzs/aops/tree/main/assets/disciplines/coordinator-loop
Command: npx skills add https://github.com/eeemzs/aops --skill coordinator-loop-eeemzs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Multi-agent sessions often drift when an operator delegates work informally: assignments get lost in chat, reviews are skipped, and handoffs lack context. This discipline defines a repeatable loop where one coordinator acts as the single operator interface and directs implementers through canonical, verifiable assignments. ## Core Features & Use Cases - Canonical assignment flow: Create or select mission, task, and sprint references before assigning work, so chat prose alone never counts as an assignment. - Mandatory review loop: Require one Projectman review request per implementation slice, verify evidence, and loop on findings until closure. - Resume-ready handoff: Close each session with a checkpoint, issue triage, review accounting, and a handoff another agent can resume from. - Use Case: An operator delegates a multi-task feature build to a coordinator agent, which assigns bounded slices to implementer agents, collects reviews, and ends with a checkpoint the team can resume the next day. ## Quick Start Use the coordinator-loop discipline to have one coordinator agent assign bounded tasks to implementers through Projectman and close the session with a checkpoint and handoff.

Frequently Asked Questions about coordinator-loop

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

FAQPage Schema
How do I coordinate multiple AI agents on one project?▼

Designate one coordinator agent as the single operator interface. The coordinator verifies each request against code and docs, creates canonical mission or task references, and assigns bounded work to implementers through Projectman rather than chat messages.

How do I assign tasks to AI agents so they are trackable?▼

Create or select canonical mission, task, or sprint references in Projectman before assigning work. Chat prose alone is not an assignment; each bounded slice must reference a canonical item and produce a review request.

Can a coordinator agent also review implementer work?▼

A coordinator may review only when independent verification is preserved. The default flow requires one Projectman review request per implementation slice, with the coordinator verifying evidence and looping on findings.

What skills does the coordinator-loop discipline require?▼

It requires aops-collaborative-work, aops-cli-projectman, aops-cli-chat, and aops-cli-agentspace. The aops-cli-discuss skill is used when material design consensus is needed before assignment.

When should I not use a coordinator-loop workflow?▼

Avoid it for single-task or solo sessions where delegation overhead adds no value. Commit, publish, and deletion actions also remain separately authorized and are not covered by the coordinator's assignment authority.