Memory

iaiops-process

Try it

Read process instrumentation across HART-IP, OPC-UA, and Modbus with built-in diagnostics for downtime, data quality, and OEE.

What it does

Connects to chemical, pharma, food & beverage, and oil & gas plants via HART-IP (transmitters, valve positioners, loop current on UDP/TCP 5094), OPC-UA (DCS and gateway read), and Modbus-TCP/RTU (skids, analyzers, serial RTU). Includes process-specific tools: PID control loop health triage (oscillation, steady-state offset, output saturation) and heat exchanger fouling detection via thermal efficiency calculation. Exposes a cross-protocol diagnostic brain covering downtime root cause analysis, data quality scoring (staleness, flatline, bad quality), OEE computation, baseline change detection, alarm bad-actor identification, and historian coverage. All protocol tools are read-only by design;…

When to use it

  • Diagnosing a HART transmitter loop fault or burst-mode communication issue
  • Pulling PV/SV/TV/QV data from a DCS via OPC-UA for OEE or anomaly analysis
  • Reading analyzer or skid registers over Modbus-TCP/RTU and mapping them to named tags
  • Running downtime root cause or data quality scorecard across mixed HART/OPC-UA/Modbus assets

The skill document

iaiops-process — 流程工业 edition(HART + Modbus + OPC-UA + 脑)

启动:IAIOPS_MCP=process / iaiops-mcp-process(= opcua + modbus + hart + 脑)。 现场有 UNS/Sparkplug 时用 IAIOPS_MCP=process,sparkplug(sparkplug 工具清单见 iaiops-factory skill)。HART 需 extra:pip install iaiops[hart]。

工具

HART-IP(只读;过程仪表 — 变送器、阀门定位器,经网关)

transport 为 udp(默认)或 tcp,端口 5094。

  • hart_device_identity — 通用设备身份(command 0)
  • hart_primary_variable — 主变量 PV(command 1):值 + 单位码
  • hart_dynamic_variables — 动态变量 PV/SV/TV/QV + 回路电流(command 3)
  • hart_burst_sample — 主动采样 burst 发布的变量;hart_burst_listen — 被动监听非请求 burst publish(待核实 per 网关)

OPC-UA(只读;DCS/网关旁路读取)

  • opcua_server_info / opcua_browse / opcua_read_node / opcua_read_many
  • opcua_subscribe_sample opcua_read_alarms opcua_alarm_events(A&C 带时间戳) opcua_read_history(HDA)
  • opcua_diagnose_connection — 连接失败归因(证书/策略/认证/网络/配置)
  • opcua_discover_tags — 自动发现 + 语义资产建模
  • opcua_health_summary — tag vs 阈值分类;opcua_anomaly_scan — 有界统计异常扫描

Modbus-TCP / Modbus-RTU(只读;撬装/分析仪/RTU 从站)

  • modbus_read_holding modbus_read_input modbus_read_coils modbus_read_discrete
  • modbus_detect_byte_order — 字节/字序自动探测
  • modbus_list_templates / modbus_apply_template — 厂商寄存器模板 → 命名 tag
  • modbus_health_summary — 寄存器 vs 阈值分类

流程专属(edition 工具;仅随 process edition 加载,不进全局脑)

  • control_loop_health — PID 回路诊断:从一段 PV/SP/OP 采样检出振荡(PV 反复穿越 SP)、 稳态偏差(PV 长期偏离 SP)、输出饱和(OP 压在 0/100%)。给 verdict(saturated>oscillating> offset>ok)。非整定器,只分诊哪些回路要看。纯分析,每项引用数值。
  • heat_exchanger_fouling — 换热器结垢检测:按四路温度算热侧温度效能 ε=(hot_in−hot_out)/ (hot_in−cold_in),前后半窗对比;效能低于阈值或下降超阈即 fouling(结垢征兆,先于强制清洗)。 纯分析,喂 OPC-UA/Modbus/HART 温度点,verdict 引用效能数值。

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(流程工业重点: staleness / flatline / bad-quality —— 仪表坏数据绝不静默插值喂 AI)
  • 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 → iaiops doctor → HART 先 hart_device_identity 证明网关链路,OPC-UA 先 opcua_diagnose_connection。
  2. Read-first:仪表读 PV(hart_primary_variable)→ 动态变量 → burst; 回路异常/告警风暴用 alarm_bad_actors(ISA-18.2),"没数据"用 diagnose_dataflow。
  3. MOC 写:本 edition 三个协议的工具全部只读(无写表面)。若叠加 sparkplug, mqtt_publish 为 [WRITE][HIGH][MOC]:默认 dry_run=True + 改前状态 undo + iaiops approve 具名审批双确认。未经授权绝不写生产控制系统。

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
HART-IPhart-protocol>=2023.6,<2025(extra)HART-IP(经网关)过程仪表:变送器/阀门定位器UDP/TCP 5094⚠️ codec CI 自测;真机网关 待核实
OPC-UAasyncua>=2.0,<3OPC UA 1.0x(DA+HA+AC 子集)任意合规 Server / DCS 网关opc.tcp✅ mock+HDA
Modbus-TCPpymodbus>=3.5,<4App 1.1b3;FC 1/2/3/4/5/6/15/16任意 TCP 从站TCP/502✅
Modbus-RTUpymodbus>=3.5,<4 + pyserial>=3.5Modbus serial (RTU)串口从站RS-485/serial✅ socat PTY verified 2026-07-02;物理 RS-485 待核实

Questions people ask

What transport layers does HART-IP use?
HART-IP operates on UDP or TCP port 5094 through a gateway. The skill provides device identity (command 0), primary variable (command 1), dynamic variables PV/SV/TV/QV with loop current (command 3), and burst sample/listen tools.
Are writes allowed to the DCS or field instruments?
No. All HART-IP, OPC-UA, and Modbus tools in this edition are strictly read-only. If Sparkplug/MQTT is enabled, mqtt_publish is gated behind MOC approval with dry_run=True and undo capability.
How does data quality handling differ from standard monitoring?
The data quality scorecard flags staleness, flatline, and bad-quality tags — ensuring corrupted or missing instrument data is not silently interpolated before feeding AI or analytics.

Related skills

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

by zw0081 installs1 stars

Measure whether a transformation changed your growth engine or just added a one-time bump.

by deciqai1 installs2 stars

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

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

by deciqai1 installs3 stars

Structured framework to analyze your alternatives and acceptable deal range before you commit

by deciqai1 installs3 stars

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

by deciqai1 installs2 stars

More from zw008

Browse all skills

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

by zw00878 installs1 stars

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

by zw00876 installs

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

by zw00853 installs

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

by zw00851 installs

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

by zw00851 installs

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

by zw00850 installs