Installable skillMIT

Resource Scheduler

Resource scheduling and assignment optimization skill for personnel and equipment allocation

BY a5c-aiUPDATED Oct 28, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: resource-scheduler-5cfe8ac3161b
description: Resource scheduling and assignment optimization skill for personnel and equipment allocation
allowed-tools:
  - Read
  - Write
  - Glob
  - Grep
  - Edit
metadata:
  specialization: operations
  domain: business
  category: workflow-automation
graph:
  domains: [domain:operations]
  skillAreas: [skill-area:capacity-planning-ops, skill-area:quantitative-modeling, skill-area:data-analytics]
  workflows: [workflow:vendor-onboarding, workflow:vendor-evaluation]
  roles: [role:operations-analyst, role:procurement-manager, role:project-manager]
---

# Resource Scheduler

## Overview

The Resource Scheduler skill provides comprehensive capabilities for optimizing resource scheduling and assignment. It supports skill-based assignment, shift scheduling, overtime optimization, and equipment allocation.

## Capabilities

- Skill-based assignment
- Shift scheduling
- Overtime optimization
- Cross-training utilization
- Equipment allocation
- Maintenance window scheduling
- Conflict resolution
- Schedule publication

## Used By Processes

- CAP-002: Production Scheduling Optimization
- CAP-001: Capacity Requirements Planning
- TOC-002: Drum-Buffer-Rope Scheduling

## Tools and Libraries

- Workforce management systems
- Scheduling optimization algorithms
- HR systems integration
- Communication platforms

## Usage

```yaml
skill: resource-scheduler
inputs:
  scheduling_horizon: 7  # days
  resources:
    - name: "John Smith"
      type: "operator"
      skills: ["assembly", "welding", "inspection"]
      shift_preference: "day"
      max_hours: 50
    - name: "Jane Doe"
      type: "operator"
      skills: ["assembly", "packaging"]
      shift_preference: "flexible"
      max_hours: 45
  requirements:
    - date: "2026-01-25"
      shift: "day"
      skill: "assembly"
      count: 3
    - date: "2026-01-25"
      shift: "day"
      skill: "welding"
      count: 2
  constraints:
    - "No consecutive night shifts"
    - "Minimum 8 hours between shifts"
    - "Maximum 10 hours per shift"
outputs:
  - schedule_assignments
  - coverage_report
  - overtime_forecast
  - skill_gaps
  - conflict_resolutions
```

## Scheduling Objectives

| Objective | Priority | Metric |
|-----------|----------|--------|
| Coverage | High | % requirements filled |
| Skill Match | High | Qualified for assignment |
| Fairness | Medium | Balanced distribution |
| Cost | Medium | Overtime minimization |
| Preference | Low | Employee satisfaction |

## Shift Patterns

| Pattern | Description | Use Case |
|---------|-------------|----------|
| Fixed | Same schedule weekly | Stable demand |
| Rotating | Shifts rotate | 24/7 operations |
| Compressed | Longer days, fewer days | Employee preference |
| Flexible | Variable start/end | Demand variation |
| Split | Two shifts per day | Peak periods |

## Skill Matrix

| Resource | Skill 1 | Skill 2 | Skill 3 |
|----------|---------|---------|---------|
| Operator A | Expert | Competent | Training |
| Operator B | Training | Expert | None |
| Operator C | Competent | Training | Expert |

## Assignment Algorithm

```
1. Identify requirements
2. Match skills to requirements
3. Apply availability constraints
4. Optimize for objectives
5. Resolve conflicts
6. Publish schedule
```

## Overtime Management

| Hours | Rate | Threshold |
|-------|------|-----------|
| 0-40 | 1.0x | Standard |
| 40-50 | 1.5x | Overtime |
| 50+ | 2.0x | Double-time |

## Cross-Training Strategy

1. Identify critical skills
2. Assess current coverage
3. Identify training candidates
4. Develop training plan
5. Track progress
6. Update skill matrix

## Integration Points

- HR/payroll systems
- Time and attendance
- ERP systems
- Communication platforms
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.