Back to home@InterPSS-Project

ipss-agent

InterPSS Agentic Power System Simulation Agent for AC load flow, DC-based contingency analysis, and NERC TPL style reporting. This repository ships agent-facing skills. It is also integrated into Deepseek Harness as a DSH Plugin. You can run power system simulations in the traditional way in a local sandbox or in the native AI env.

Stars
2
Language
Python
Created
May 11, 2026
Updated
Aug 20, 2026

Introduction

iPSS Agent

InterPSS Agentic Power System Simulation Agent for AC load flow, DC-based contingency analysis, and NERC TPL-001-5 style reporting. This repository ships agent-facing skills for OpenAI Codex Desktop and Claude Code CLI (see Setup.md). Other agent CLIs can reuse the same workflow by copying the skill content from .agents/skills/ipss-sim/ or .claude/skills/ipss-sim/.

It is also integrated into Deepseek Harness as a DSH Plugin. You can run power system simulation in the traditional step-by-step way:

ipss-dsh-plugin

Or in the native AI Chat way:

ipss-agent-chat

Environment setup

Git check out or download this repository and follow the setup instructions below.

OpenAI Codex Desktop (recommended)

  1. Add the folder as a Codex project.
  2. Run the setup below, then test with a sample case (for example, under wspace/data/psse/Texas2K).
Setup ipss.agent env
# Test the setup with a sample case directory
/ipss-sim <simu_case_directory> "<NERC Report Name>"

After completing the env setup, add your data folder as a Codex project and use the agent to do the simulations.

Claude Code CLI setup

/init  # optional
Setup ipss.agent env
Setup the skills and commands to run the workflow

DSH Plugin setup

Download the DSH Plugin from the Release page and install it following the provided instructions.

Run simulations and generate reports

After the setup, run(Chat) power-system simulations through the agent.

Skill-style simulation

Input file mode:

/ipss-sim <simu_case_file> <contingency_file> <monitored_file> "<NERC Report Name>"

Input directory mode (auto-discovers *.RAW / *.raw / *.ieee, *contingency*.json, *monitor*.json):

/ipss-sim <simu_case_directory> "<NERC Report Name>"
/ipss-sim Aclf only <simu_case_directory> "<Loadflow Report Name>"

/NERC Report HTML         # generate Html report
/NERC Report Slides       # generate PPT report

Direct prompt simulation

Run ACLF format psse <case_file>
Run CA format psse <case_file> <contingency_file> <monitor_file>
Gen NERC TPL report "<NERC Report Name>" <result_dir>

Explore simulation results

Results are written under wspace/<parent_of_case_file>/result/ (the directory that contains your case file, plus a result subfolder). You can inspect them with the LLM, for example:

# Load flow
Find the lowest voltage bus
Find the highest loading branch
# Contingency analysis
Find the top N-1 loaded branches

Reference

TopicDocument
Layout, JARs, config.json, JVMSetup.md
src/ipss_cmd.py usageIpssCmd.md
Report generatorGenReport.md