AgentSkillsCN

eoa-module-management-commands

在模块管理任务中使用。当需要在编排阶段对模块进行增删改、调整优先级或重新分配时,可调用此技能。每个模块均与 GitHub Issue 实现一一映射。

SKILL.md
--- frontmatter
name: eoa-module-management-commands
description: "Trigger with module management tasks. Use when managing modules during Orchestration Phase (add, modify, remove, prioritize, reassign). Every module maps 1:1 to GitHub Issue."
license: Apache-2.0
compatibility: Cross-platform compatible. Requires Python 3.8+ and PyYAML for scripts. Requires gh CLI for GitHub Issue operations. Works with AI Maestro messaging for agent notifications. Requires AI Maestro installed.
metadata:
  author: Anthropic
  version: 1.0.0
user-invocable: false
context: fork
agent: eoa-main

Module Management Commands Skill

Overview

This skill teaches orchestrators how to dynamically manage modules during the Orchestration Phase.

Prerequisites

  • Orchestration Phase active (Plan Phase completed and approved)
  • GitHub CLI (gh) authenticated
  • AI Maestro running for agent notifications
  • State file design/state/exec-phase.md exists Modules are the atomic units of work in Atlas orchestration. Each module represents one feature, component, or deliverable that an agent will implement.

CRITICAL RULE: Every module is tied 1:1 to a GitHub Issue. When you add a module, an issue is created. When you remove a module, the issue is closed. When you modify a module, the issue is updated. This linkage ensures traceability and transparency.

Instructions

  1. Identify the module management action needed (add, modify, remove, prioritize, or reassign)
  2. Verify prerequisites: Orchestration Phase active, gh CLI authenticated, AI Maestro running
  3. Execute the appropriate command with required parameters
  4. Verify the module state update in design/state/exec-phase.md
  5. Confirm GitHub Issue synchronization completed
  6. Notify affected agents via AI Maestro if applicable

Output

Output TypeFormatLocation
Module state updateYAMLdesign/state/exec-phase.md
GitHub Issue created/updatedIssue numberGitHub repository
Agent notificationAI Maestro messageAgent inbox
Command success/failureStatus messageConsole output

Commands Overview

CommandPurposeWhen to Use
/add-moduleAdd new moduleUser requests new feature
/modify-moduleChange module specsCriteria or priority change
/remove-moduleDelete pending moduleModule cancelled
/prioritize-moduleChange priority levelUrgency changed
/reassign-moduleTransfer to different agentAgent stuck or unavailable

Quick Reference by Situation

Situation: User wants a new feature

  1. Use /add-module with name and criteria
  2. System creates GitHub Issue automatically
  3. Module appears in pending queue
  4. Assign to agent when ready

Read: module-creation.md

  • 1.1 When to add modules during orchestration
  • 1.2 Required fields for new modules (name, criteria)
  • 1.3 Optional fields (priority level)
  • 1.4 Automatic GitHub Issue creation
  • 1.5 State file update after addition
  • 1.6 Complete examples with all variations

Situation: Requirements changed for a module

  1. Use /modify-module with new specs
  2. System updates linked GitHub Issue
  3. If module assigned, agent is notified

Read: module-modification.md

  • 2.1 What can be modified (name, criteria, priority)
  • 2.2 Modification restrictions by status
  • 2.3 Agent notification protocol
  • 2.4 GitHub Issue synchronization
  • 2.5 Complete modification examples

Situation: Module needs to be cancelled

  1. Verify module status is pending
  2. Use /remove-module with module ID
  3. System closes linked GitHub Issue

Read: module-removal-rules.md

  • 3.1 Which modules can be removed (pending only)
  • 3.2 Why in-progress modules cannot be removed
  • 3.3 Removal process step by step
  • 3.4 GitHub Issue closure with wontfix label
  • 3.5 Alternatives to removal (scope reduction)
  • 3.6 Error handling and recovery

Situation: Module urgency changed

  1. Use /prioritize-module with new level
  2. System updates GitHub Issue labels
  3. Agent notified if assigned

Read: module-prioritization.md

  • 4.1 Priority levels explained (critical, high, medium, low)
  • 4.2 Effects on assignment queue
  • 4.3 GitHub Issue label updates
  • 4.4 When to escalate vs downgrade
  • 4.5 Complete priority change examples

Situation: Module needs different agent

  1. Request progress report from current agent
  2. Use /reassign-module with new agent ID
  3. Old agent notified to stop work
  4. New agent receives full assignment

Read: module-reassignment.md

  • 5.1 When reassignment is appropriate
  • 5.2 Reassignment workflow step by step
  • 5.3 Old agent notification protocol
  • 5.4 New agent assignment message
  • 5.5 State file updates during reassignment
  • 5.6 Instruction Verification Protocol reset

The Module-Issue Relationship

Every module in Atlas orchestration has a corresponding GitHub Issue. This relationship is fundamental and cannot be bypassed.

Why Modules Map to Issues

ReasonBenefit
TraceabilityAll work is tracked in GitHub
TransparencyStakeholders see progress
IntegrationPRs link to issues
HistoryChanges documented
AccountabilityAssignment visible

Automatic Synchronization

Module EventGitHub Issue Action
Module addedIssue created with labels
Module modifiedIssue body/labels updated
Module removedIssue closed with wontfix
Priority changedPriority label updated
Agent assignedIssue assigned to developer
Module completedIssue closed with linked PR

Read: github-issue-sync.md

  • 6.1 Issue creation format and labels
  • 6.2 Issue update synchronization
  • 6.3 Issue closure protocols
  • 6.4 Label conventions (module, priority-, status-)
  • 6.5 Manual sync when automation fails
  • 6.6 Troubleshooting sync issues

Command Details

/add-module

Usage: /add-module "<NAME>" --criteria "<TEXT>" [--priority LEVEL]

ArgumentRequiredDescription
NAMEYesDisplay name for the module
--criteriaYesAcceptance criteria text
--priorityNocritical, high, medium, low (default: medium)

What Happens:

  1. Module entry created in state file
  2. GitHub Issue created with labels
  3. Stop hook updated to include new module

Example:

bash
/add-module "Two-Factor Auth" --criteria "Support TOTP and SMS" --priority critical

/modify-module

Usage: /modify-module <MODULE_ID> [--name NAME] [--criteria TEXT] [--priority LEVEL]

ArgumentRequiredDescription
MODULE_IDYesID of module to modify
--nameNoNew display name
--criteriaNoNew acceptance criteria
--priorityNoNew priority level

Restrictions:

  • pending modules: All fields modifiable
  • in_progress modules: Modifiable with agent notification
  • complete modules: Cannot modify

Example:

bash
/modify-module auth-core --criteria "Support JWT with 24h expiry" --priority high

/remove-module

Usage: /remove-module <MODULE_ID> [--force]

ArgumentRequiredDescription
MODULE_IDYesID of module to remove
--forceNoSkip confirmation

Restrictions:

  • Only pending modules can be removed
  • in_progress modules cannot be removed
  • complete modules cannot be removed

Example:

bash
/remove-module oauth-facebook

/prioritize-module

Usage: /prioritize-module <MODULE_ID> --priority <LEVEL>

ArgumentRequiredDescription
MODULE_IDYesID of module
--priorityYescritical, high, medium, low

Example:

bash
/prioritize-module auth-core --priority critical

/reassign-module

Usage: /reassign-module <MODULE_ID> --to <AGENT_ID>

ArgumentRequiredDescription
MODULE_IDYesID of module to reassign
--toYesID of new agent

What Happens:

  1. Old agent receives STOP notification
  2. Assignment record transferred
  3. New agent receives full assignment
  4. Instruction Verification Protocol resets

Example:

bash
/reassign-module auth-core --to implementer-2

State File Structure

Module management commands modify the orchestration state file at design/state/exec-phase.md.

Module Entry Format

yaml
modules_status:
  - id: "auth-core"           # Kebab-case identifier
    name: "Core Authentication"  # Display name
    status: "pending"         # pending|assigned|in_progress|complete
    assigned_to: null         # Agent ID or null
    github_issue: "#42"       # Linked issue number
    pr: null                  # Linked PR when complete
    verification_loops: 0     # Number of review cycles
    acceptance_criteria: "Support JWT and session tokens"
    priority: "high"          # critical|high|medium|low

Assignment Entry Format

yaml
active_assignments:
  - agent: "implementer-1"
    agent_type: "ai"
    module: "auth-core"
    github_issue: "#42"
    task_uuid: "task-abc123def456"
    status: "pending_verification"
    instruction_verification:
      status: "awaiting_repetition"
      repetition_received: false
      repetition_correct: false
      questions_asked: 0
      questions_answered: 0
      authorized_at: null

Examples

Example 1: Add New Module Mid-Orchestration

bash
# User requests two-factor authentication
/add-module "Two-Factor Auth" --criteria "Support TOTP and SMS" --priority critical

# System automatically creates GitHub Issue #43
# Module appears in pending queue with status: pending

Example 2: Reassign Blocked Module

bash
# Agent implementer-1 is blocked on auth-core
# Request progress report first
/check-agents --agent implementer-1

# Reassign to implementer-2
/reassign-module auth-core --to implementer-2

# Old agent notified to stop
# New agent receives full assignment with Instruction Verification Protocol reset

Error Handling

ProblemLikely CauseSolution
Module not foundWrong ID usedUse /orchestration-status to see actual IDs
Cannot remove moduleStatus not pendingModify scope instead of removing
GitHub Issue not createdgh CLI not authenticatedRun gh auth login
Agent not notifiedAI Maestro not runningCheck http://localhost:23000
Priority not updatingIssue labels mismatchManually update GitHub Issue

For detailed troubleshooting: troubleshooting.md

  • 7.1 State file corruption recovery
  • 7.2 GitHub sync failure recovery
  • 7.3 Agent notification failures
  • 7.4 Module ID conflicts
  • 7.5 Force removal scenarios

Scripts Reference

Module Operations Script

Location: scripts/module_operations.py

Provides programmatic access to all module operations:

bash
# Add module
python3 scripts/module_operations.py add "Module Name" --criteria "Criteria"

# Modify module
python3 scripts/module_operations.py modify module-id --priority critical

# Remove module
python3 scripts/module_operations.py remove module-id

# List modules
python3 scripts/module_operations.py list

# Validate module state
python3 scripts/module_operations.py validate

GitHub Sync Script

Location: scripts/github_sync.py

Synchronizes module state with GitHub Issues:

bash
# Sync all modules to GitHub
python3 scripts/github_sync.py sync-all

# Sync specific module
python3 scripts/github_sync.py sync module-id

# Verify sync status
python3 scripts/github_sync.py verify

Reference Documents

ReferenceContents
module-creation.mdAdd-module workflow, validation, examples
module-modification.mdModify specs, notifications, restrictions
module-removal-rules.mdRemoval conditions, cleanup, alternatives
module-prioritization.mdPriority levels, effects, label updates
module-reassignment.mdTransfer workflow, notifications, reset
github-issue-sync.mdIssue creation, labels, sync protocol
troubleshooting.mdError recovery, force operations

Related Commands

CommandPurpose
/orchestration-statusView all modules and assignments
/assign-moduleInitial assignment to agent
/check-agentsMonitor agent progress
/register-agentRegister new agents

Summary

ActionCommandRestrictions
Add module/add-moduleNone
Modify specs/modify-moduleCannot modify complete
Remove module/remove-modulePending only
Change priority/prioritize-moduleNone
Reassign agent/reassign-moduleCannot reassign complete

Key Principles:

  • Every module = 1 GitHub Issue
  • Modifications sync to GitHub automatically
  • Agent notifications via AI Maestro
  • Instruction Verification required after reassignment

Resources


Checklist

Copy this checklist and track your progress:

  • Identify module management action needed (add/modify/remove/prioritize/reassign)
  • Verify Orchestration Phase is active
  • Confirm gh CLI is authenticated
  • Verify AI Maestro is running (http://localhost:23000)
  • Execute appropriate command with required parameters
  • Check module state updated in design/state/exec-phase.md
  • Verify GitHub Issue created/updated with correct labels
  • Confirm agent notification sent via AI Maestro (if applicable)
  • Validate module-to-issue 1:1 relationship maintained
  • Document any errors in troubleshooting log