Data & analysis

iaiops-fab

Try it

SECS/GEM + OPC-UA fab equipment monitoring with cross-protocol downtime and OEE intelligence

What it does

Connects to semiconductor and display fab equipment over SECS/GEM (SEMI E5, E30, E37 HSMS) for MES interfaces and OPC-UA for internal PLC control layers. Reads status variables (SVID), equipment constants (ECID), alarms (ALID), and process programs (PPID) from the SECS/GEM side; browses nodes, reads values, subscribes, and queries history from the OPC-UA side. Includes fab-specific analysis tools: statistical process control (Western Electric/Nelson rules with Cp/Cpk), defect Pareto charts, OEE computation, downtime root-cause copilot, data quality scoring, and cross-protocol asset inventory. All SECS/GEM and OPC-UA tools are read-only by design. Write operations are gated behind MOC approv…

When to use it

  • Investigating equipment downtime using SECS/GEM status variables and OPC-UA internal signals together
  • Running SPC checks on wafer/panel measurement sequences to detect out-of-control conditions
  • Generating defect Pareto charts to identify the vital few defect categories
  • Auditing data quality and tag health across a fab equipment fleet via cross-protocol asset model

The skill document

iaiops-fab — 半导体/显示 fab edition（SECS/GEM + OPC-UA + 脑）

启动：IAIOPS_MCP=fab / iaiops-mcp-fab（暴露 secsgem + opcua + s7 + modbus + 脑）。 安装：pip install iaiops[fab]。我们是 HOST（HSMS ACTIVE），设备是 equipment。

Fab 分层认知：一台 fab 设备是两层 —— MES-facing 的 SECS/GEM(HSMS) 接口，与 设备内部控制（PLC，走 OPC-UA/S7/Modbus）。职责不同，别混。S7/Modbus 工具清单见 iaiops-factory skill（fab profile 已同时暴露）。

工具

SECS/GEM（只读；设备 ↔ MES 标准，fab 入场券）

  • secsgem_equipment_status — 建立 GEM 链路 + Are-You-There（S1F1/F2）
  • secsgem_list_status_variables — SVID namelist（S1F11/F12）
  • secsgem_read_status_variables — SVID 值（S1F3/F4）
  • secsgem_list_equipment_constants — ECID namelist（S2F29/F30）
  • secsgem_read_equipment_constants — ECID 值（S2F13/F14）
  • secsgem_list_alarms — 告警表（S5F5/F6）：ALID、ALCD、text
  • secsgem_list_process_programs — PPID 目录（S7F19/F20）

OPC-UA（只读；设备内控层 / opc.tcp 端点）

  • opcua_server_info / opcua_browse / opcua_read_node / opcua_read_many
  • opcua_subscribe_sample — 有界采样后返回（绝不死循环）
  • opcua_read_alarms — best-effort 活动告警/condition（untimed）；opcua_alarm_events — A&C 事件订阅（带服务器时间戳）
  • opcua_read_history — HDA 历史读取（[start,end] 窗口）
  • opcua_diagnose_connection — 连不上时归因（证书/安全策略/认证/防火墙/DNS/端口/配置）
  • opcua_discover_tags — 自动发现 tag 并建语义资产模型
  • opcua_health_summary — tag vs 阈值分类；opcua_anomaly_scan — 有界统计异常扫描

Fab / 质量专属（edition 工具;仅随 fab edition 加载,不进全局脑）

  • spc_check — 统计过程控制:对一段量测序列套用 Western Electric / Nelson 控制图规则 (越 3σ、2/3 越 2σ、4/5 越 1σ、连续 8 点单侧、6 点趋势),逐条按点索引报违规;给 USL/LSL 则附 Cp/Cpk。判 in_control / out_of_control。纯分析,每违规引用触发点索引。
  • defect_pareto — 缺陷帕累托:按缺陷类别计数排序,算各类占比与累计占比,标出到 80% 线的 关键少数(遏制/改善最有杠杆的类别)。纯分析,喂检验/缺陷记录,每占比引用计数。

跨协议脑（永远随 server 暴露）

  • 诊断：diagnose_dataflow downtime_root_cause downtime_root_cause_live downtime_triage learn_cause_weights rca_corpus_from_maintenance historian_health alarm_bad_actors tag_health subscription_health heartbeat_health alarm_flood_analysis alarm_cascade alarm_rationalization_worksheet
  • 数据质量：data_quality_scorecard data_quality_fleet_rollup
  • 分析：oee_compute downtime_events oee_multidim monitor_changes health_summary (deprecated) anomaly_scan (deprecated)
  • 资产：asset_inventory cross_protocol_asset_model adopt_alias_map diff_alias_map
  • 基线：baseline_learn baseline_check baseline_record_change baseline_status （change-log 基线：拒学薄历史、只报持续越带、每次告警必引基线样本 —— 非黑盒异常检测）
  • 合规/信创：compliance_mapping compliance_frameworks compliance_dengbao_levels compliance_report compliance_evidence_bundle historian_push export_data historian_query historian_coverage stream_publish stream_publish_event rca_narrate fleet_status fleet_incidents pdm_forecast historian_push export_data
  • 程序解读：plc_program_outline plc_program_xref plc_program_section plc_program_visibility（解读导出的 ST/AWL/L5X 程序,只读文件,强制引用行号）
  • 元：protocols_supported

Workflows

  1. Doctor-first：protocols_supported 看配置 → iaiops doctor 探活 → secsgem_equipment_status 确认 GEM 链路（S1F1/F2 通了再谈其他）。
  2. Read-first：SVID/ECID namelist → 值 → secsgem_list_alarms；PLC 层用 OPC-UA 工具；停机问题直接 downtime_root_cause_live（advisory，引用真实信号）。
  3. MOC 写：本 edition 的 SECS/GEM 与 OPC-UA 工具均只读；fab profile 暴露的写工具 （如 s7_write_db）走统一 MOC：risk=HIGH + 默认 dry_run=True + 改前值 undo + 具名审批人 iaiops approve 双确认。未经授权绝不写生产设备。

支持版本矩阵（内部 HLD §8，设计文档不随本仓发布；待核实 不得当既成事实）

协议库(pin)规范/版本覆盖传输自测
SECS/GEMsecsgem>=0.3,<1SECS-II(E5) · GEM(E30) · HSMS(E37/TCP)；SECS-I(E4) 待核实面板/半导体设备 ↔ MES（HOST 侧）HSMS/TCP✅ host+equipment 全软件自测；真机 待核实
OPC-UAasyncua>=2.0,<3OPC UA 1.0x（DA+HA+AC 子集）；FX/TSN 在路线任意合规 Serveropc.tcp✅ mock+HDA

Questions people ask

Does this skill write to equipment, or is it read-only?
The SECS/GEM and OPC-UA tool sets are read-only. Write tools exposed by the fab profile require MOC approval with explicit risk=HIGH, dry_run=True, and dual confirmation before any production write.
What SECS/GEM message streams does this cover?
The document specifies S1 (establish connection, Are-You-There), S2 (equipment constants), S5 (alarms), S7 (process programs). Full GEM state model and event reporting are part of the E30 specification coverage but verify against your equipment's implementation.
Can it analyze PLC programs from S7 or Modbus equipment?
PLC program interpretation tools (plc_program_outline, plc_program_xref, plc_program_section) read exported ST/AWL/L5X files and cross-reference by line number. S7/Modbus tools themselves are documented under the iaiops-factory skill; the fab profile exposes both.

Related skills

k8s-aiops

Operate Kubernetes clusters with 55 audited tools — list resources, diagnose pod health, scale workloads, and manage rollouts safely.

by zw0081 installs1 stars

Cognitive Science Landscape

Diagnose which mental domain is holding you back before choosing a cognitive intervention.

by deciqai1 installs3 stars

Feynman Technique

Tests whether you genuinely understand something or just recognize it — exposes the gaps in your mental model.

by deciqai1 installs2 stars

unusual-options-activity

End-of-day options analytics ranked against each ticker's own history: IV rank, put/call percentile, skew, max pain, and unusually active contracts.

by thesentitrader2 installs2 stars

Tunnel Vision & Slack

Escape the scarcity trap — diagnose bandwidth consumption and design protected slack to restore strategic capacity.

by deciqai1 installs2 stars

Ansoff Matrix

Prioritize growth directions with a 2×2 risk framework — pick one bet and commit.

by deciqai2 installs2 stars

More from zw008

Browse all skills

vmware-aiops

Operate VMware VMs, deployments, clusters, guest tasks, and alarms with plan and rollback support.

by zw00878 installs1 stars

vmware-monitor

Inspect VMware health, inventory, alarms, events, and performance without changing infrastructure.

by zw00876 installs

vmware-aria

Query Aria Operations metrics, alerts, capacity forecasts, anomalies, and reports from CLI or MCP.

by zw00853 installs

vmware-avi

Manage AVI services and pools, and diagnose AKO ingress, sync, certificates, analytics, and health.

by zw00851 installs

vmware-vks

Manage Supervisor Namespaces and TKC cluster lifecycles in vSphere Kubernetes Service.

by zw00851 installs

vmware-nsx

Manage NSX segments, gateways, routing, IP pools, health checks, and connectivity diagnostics.

by zw00850 installs