用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/a5c-ai/babysitter --skill line-balancer命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Reference for querying the Atlas knowledge graph through its MCP tools — the SECONDARY enrichment/comparison layer that adds best-practice context to systems you have ALREADY scanned from your real sources (`az`, repos, dirs). Use when you need to look up nodes, edges, kinds, clusters, stats, or wiki pages in Atlas to compare against your real inventory. (atlas graph, query atlas, atlas mcp, search the graph, graph neighbors, atlas record, atlas kinds, enrichment layer)
Atlas turns your STATED NEED into a real systems atlas by SCANNING your actual sources (Azure via `az`, git repos, local dirs) and process/data mining them, THEN enriching against the Atlas knowledge graph. Use this skill when asked to inventory/map your real systems, scan your cloud + repos + directories, mine the real processes or data they contain, or collect their real constraints/gotchas. (atlas, scan my systems, inventory our azure account, map my repos, real systems atlas, process mining, data mining, collect nuances, system discovery)
This skill should be used when the user asks to "find skills in the wild", "assimilate popular workflows", "discover SKILL.md files in repos", "research external skills", "find workflow patterns", "survey the skill landscape", "what skills exist out there", or wants to investigate public repositories for extractable processes, babysitter plugins, and reusable procedural insights. Searches GitHub for SKILL.md files, classifies repos by archetype, and maintains structured research under docs/reference-repos/.
基于 SOC 职业分类
正在显示 SKILL.md
| name | line-balancer |
| description | Assembly line balancing skill for workstation design and cycle time optimization. |
| allowed-tools | Bash(*) Read Write Edit Glob Grep WebFetch |
| metadata | {"author":"babysitter-sdk","version":"1.0.0","category":"production-planning","backlog-id":"SK-IE-030"} |
| graph | {"domains":["domain:industrial-engineering"],"skillAreas":["skill-area:statistical-analysis","skill-area:organizational-design","skill-area:data-analysis"],"roles":["role:operations-analyst","role:research-engineer"]} |
You are line-balancer - a specialized skill for assembly line balancing including workstation design, task assignment, and cycle time optimization.
This skill enables AI-powered line balancing including:
import networkx as nx
import pandas as pd
from collections import defaultdict
def analyze_precedence(tasks: list, precedence: list):
"""
Analyze precedence relationships for line balancing
tasks: list of {'task_id': str, 'time': float, 'description': str}
precedence: list of (predecessor, successor) tuples
"""
# Build directed graph
G = nx.DiGraph()
task_dict = {t['task_id']: t for t in tasks}
for task in tasks:
G.add_node(task['task_id'], time=task['time'])
for pred, succ in precedence:
G.add_edge(pred, succ)
# Calculate position weights (sum of task time and all successors)
def positional_weight(node):
descendants = nx.descendants(G, node)
weight = task_dict[node]['time']
for d in descendants:
weight += task_dict[d]['time']
return weight
weights = {t['task_id']: positional_weight(t['task_id']) for t in tasks}
# Find critical path
total_time = sum(t['time'] for t in tasks)
# Find immediate predecessors and successors
analysis = []
for task in tasks:
tid = task['task_id']
analysis.append({
'task_id': tid,
'time': task['time'],
'predecessors': list(G.predecessors(tid)),
'successors': list(G.successors(tid)),
'positional_weight': weights[tid]
})
return {
"total_work_content": total_time,
"task_analysis": pd.DataFrame(analysis).sort_values('positional_weight', ascending=False),
"graph": G
}
def calculate_cycle_time(demand_per_shift: int, available_time_minutes: float,
efficiency: float = 0.95):
"""
Calculate required cycle time from demand
Returns theoretical and practical cycle times
"""
# Theoretical cycle time
theoretical_ct = available_time_minutes / demand_per_shift
# Practical cycle time (accounting for efficiency)
practical_ct = theoretical_ct * efficiency
return {
"theoretical_cycle_time": round(theoretical_ct, 2),
"practical_cycle_time": round(practical_ct, 2),
"demand_per_shift": demand_per_shift,
"available_time": available_time_minutes,
"efficiency_factor": efficiency
}
def calculate_workstations(total_work_content: float, cycle_time: float):
"""
Calculate theoretical and actual number of workstations
"""
theoretical = total_work_content / cycle_time
minimum = int(np.ceil(theoretical))
return {
"theoretical_workstations": round(theoretical, 2),
"minimum_workstations": minimum,
"total_work_content": total_work_content,
"cycle_time": cycle_time
}
def largest_candidate_rule(tasks: list, precedence: list, cycle_time: float):
"""
Line balancing using Largest Candidate Rule
Assigns tasks to workstations by largest task time first
"""
# Build precedence graph
G = nx.DiGraph()
for pred, succ in precedence:
G.add_edge(pred, succ)
task_dict = {t['task_id']: t['time'] for t in tasks}
# Sort tasks by time descending
sorted_tasks = sorted(tasks, key=lambda x: x['time'], reverse=True)
workstations = []
assigned = set()
current_station = 1
current_time = 0
current_tasks = []
while len(assigned) < len(tasks):
task_assigned = False
for task in sorted_tasks:
tid = task['task_id']
if tid in assigned:
continue
# Check precedence - all predecessors must be assigned
predecessors = set(G.predecessors(tid))
if not predecessors.issubset(assigned):
continue
# Check if task fits in current station
if current_time + task['time'] <= cycle_time:
current_tasks.append(tid)
current_time += task['time']
assigned.add(tid)
task_assigned =
task_assigned:
current_tasks:
workstations.append({
: current_station,
: current_tasks,
: current_time,
: cycle_time - current_time
})
current_station +=
current_time =
current_tasks = []
current_tasks:
workstations.append({
: current_station,
: current_tasks,
: current_time,
: cycle_time - current_time
})
{
: workstations,
: (workstations),
: cycle_time
}
def ranked_positional_weight(tasks: list, precedence: list, cycle_time: float):
"""
Line balancing using Ranked Positional Weight method
Better than LCR as it considers both task time and position
"""
# Build graph and calculate positional weights
G = nx.DiGraph()
for pred, succ in precedence:
G.add_edge(pred, succ)
task_dict = {t['task_id']: t for t in tasks}
def calc_rpw(task_id):
descendants = nx.descendants(G, task_id)
weight = task_dict[task_id]['time']
for d in descendants:
weight += task_dict[d]['time']
return weight
# Add RPW to tasks and sort
for task in tasks:
task['rpw'] = calc_rpw(task['task_id'])
sorted_tasks = sorted(tasks, key=lambda x: x['rpw'], reverse=True)
# Assign to workstations
workstations = []
assigned = set()
current_station = 1
current_time = 0
current_tasks = []
while len(assigned) < len(tasks):
task_assigned = False
for task in sorted_tasks:
tid = task['task_id']
if tid in assigned:
predecessors = (G.predecessors(tid))
predecessors.issubset(assigned):
current_time + task[] <= cycle_time:
current_tasks.append({
: tid,
: task[],
: task[]
})
current_time += task[]
assigned.add(tid)
task_assigned =
task_assigned:
current_tasks:
workstations.append({
: current_station,
: current_tasks,
: current_time,
: cycle_time - current_time,
: current_time / cycle_time *
})
current_station +=
current_time =
current_tasks = []
current_tasks:
workstations.append({
: current_station,
: current_tasks,
: current_time,
: cycle_time - current_time,
: current_time / cycle_time *
})
{
: workstations,
: (workstations),
: cycle_time,
:
}
def calculate_line_efficiency(workstations: list, cycle_time: float, total_work_content: float):
"""
Calculate line balancing efficiency metrics
"""
num_stations = len(workstations)
# Line efficiency (balance efficiency)
line_efficiency = (total_work_content / (num_stations * cycle_time)) * 100
# Balance delay
balance_delay = 100 - line_efficiency
# Smoothness index
station_times = [ws['total_time'] for ws in workstations]
mean_time = np.mean(station_times)
smoothness = np.sqrt(sum((t - mean_time)**2 for t in station_times))
# Station utilization
utilizations = [ws['total_time'] / cycle_time * 100 for ws in workstations]
return {
"line_efficiency": round(line_efficiency, 2),
"balance_delay": round(balance_delay, 2),
"smoothness_index": round(smoothness, 2),
"num_stations": num_stations,
"cycle_time": cycle_time,
"station_utilizations": utilizations,
"min_utilization": round(min(utilizations), 2),
"max_utilization": round((utilizations), ),
: (np.mean(utilizations), )
}
def mixed_model_balance(models: list, tasks: dict, precedence: dict,
demand_ratio: dict, cycle_time: float):
"""
Balance a mixed-model assembly line
models: list of model IDs
tasks: {model: [{'task_id': str, 'time': float}]}
precedence: {model: [(pred, succ)]}
demand_ratio: {model: proportion of demand}
"""
# Calculate weighted average task times
weighted_tasks = defaultdict(float)
for model in models:
ratio = demand_ratio[model]
for task in tasks[model]:
weighted_tasks[task['task_id']] += task['time'] * ratio
# Create combined task list
combined_tasks = [
{'task_id': tid, 'time': time}
for tid, time in weighted_tasks.items()
]
# Combine precedence relationships
combined_precedence = set()
for model in models:
for pred, succ in precedence[model]:
combined_precedence.add((pred, succ))
# Balance using weighted times
result = ranked_positional_weight(
combined_tasks,
list(combined_precedence),
cycle_time
)
return {
"mixed_model_balance": result,
"models": models,
"demand_ratios": demand_ratio,
"weighted_work_content": sum(weighted_tasks.values())
}
This skill integrates with the following processes:
assembly-line-design.jsproduction-scheduling-optimization.jsworkstation-design-optimization.js{
"line_balance": {
"workstations": [
{"station": 1, "tasks": ["A", "B"], "total_time": 48, "idle_time": 2},
{"station": 2, "tasks": ["C", "D", "E"], "total_time": 47, "idle_time": 3}
],
"cycle_time": 50,
"method":