Opp repl overview

Skill tabgab/opp_repl-skill/opp-repl-overview

START HERE. Entry-point skill for opp_repl — the Python REPL, CLI tools, and MCP server for OMNeT++ simulations (github.com/omnetpp/opp_repl). Contains a decision tree that maps the user's task to the exact sibling skills to load. Load this first whenever the user asks about opp_repl, OMNeT++ simulations, running sims from Python, or regression-testing simulation models. Even weak models can navigate the rest of the pack via this skill's triage table.From its SKILL.md

Install

npx -y skills add tabgab/opp_repl-skill --skill opp-repl-overview

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

8.2 KB, ~2.0k tokens by cl100k_base, as published. Nobody here has run it

opp_repl — overview and skill map

opp_repl replaces ad-hoc shell loops and Qtenv clicking with reproducible, scriptable workflows for OMNeT++. Same feature set, three interfaces:

An IPython REPL (opp_repl) — interactive Python with all helpers pre-loaded.
A shell wrapper suite (opp_build_project, opp_run_simulations, opp_run_*_tests, opp_update_*_test_results) — for CI and scripts.
An MCP server — so AI agents can drive the REPL as a tool endpoint (a Unix-socket stdio bridge for local clients like Claude Code, or a TCP+token endpoint for remote ones).

Decision tree — load exactly these skills for common tasks

I want to...

...install opp_repl. -> opp-repl-installation

...create a NEW OMNeT++ project from scratch (write files, build, run). -> opp-repl-project-scaffolding (← START HERE for "make a new simulation", "create a new model", "build an example from zero"). On current opp_repl this is ONE function call: create_project(name, path=..., namespace=False). -> then opp-repl-running-simulations -> then opp-repl-result-analysis to read the numbers out (on current opp_repl: r.get_scalars() on the result). -> keep opp-repl-troubleshooting on hand; new projects hit §4 of that skill on namespace/package mismatches.

...run simulations that already exist. -> opp-repl-running-simulations + opp-repl-filtering -> opp-repl-tasks-and-results for drill-down / rerun -> opp-repl-result-analysis for reading the numbers.

...regression-test my project. Pick the type:

opp-repl-smoke-tests — it ran, it didn't crash.
opp-repl-fingerprint-tests — behavior hasn't drifted.
opp-repl-statistical-tests — scalar results in tolerance.
opp-repl-speed-tests — CPU instr. count stable.
opp-repl-chart-tests — analysis charts unchanged.
opp-repl-module-image-tests — module/network rendering unchanged.
opp-repl-sanitizer-tests — ASAN/UBSan clean.
opp-repl-feature-and-release-tests — comprehensive suites. All test skills cross-reference opp-repl-running-simulations and opp-repl-tasks-and-results.

...compare / bisect / debug a regression. -> opp-repl-comparing-simulations + opp-repl-fingerprint-tests -> opp-repl-tasks-and-results for post-mortem inspection.

...tune a parameter to match a target. -> opp-repl-parameter-optimization + opp-repl-running-simulations.

...measure speed / cost. -> opp-repl-profiling (call-stack profiles) or opp-repl-speed-tests (regression) or opp-repl-coverage-reports (line coverage).

...distribute runs across multiple machines. -> opp-repl-ssh-cluster + opp-repl-running-simulations.

...read .sca / .vec result files after a run. -> opp-repl-result-analysis (includes a bundled script).

...debug a build or run failure. -> opp-repl-troubleshooting FIRST — it maps every common symptom ("Building X failed", exit 127, "Class not found", etc.) to its exact cause and fix.

...wire opp_repl into an AI agent / MCP client. -> opp-repl-mcp-server (stdio bridge over a Unix socket, or TCP+token) + opp-repl-ai-workflows.

...let a human (in a terminal) and the AI share ONE live REPL. -> opp-repl-shared-terminal (tmux + MCP, same namespace).

...isolate the agent's execute_python under bubblewrap. -> opp-repl-sandbox.

...set up CI on GitHub. -> opp-repl-cli-tools + opp-repl-github-actions.

...use a specific opp_env-managed OMNeT++ version (e.g. omnetpp-6.3.0 + inet-4.6.0). -> opp-repl-opp-env-integration + opp-repl-opp-files.

Compact feature inventory

Capability	Skill
Install / setenv	`opp-repl-installation`
Core object model	`opp-repl-concepts`
`.opp` descriptors	`opp-repl-opp-files`
NEW project from zero	`opp-repl-project-scaffolding`
Interactive REPL usage	`opp-repl-repl-usage`
Shell wrappers for CI	`opp-repl-cli-tools`
Build + run simulations	`opp-repl-running-simulations`
Filtering configs/runs	`opp-repl-filtering`
Task & result model	`opp-repl-tasks-and-results`
Read .sca / .vec	`opp-repl-result-analysis`
Error decoder	`opp-repl-troubleshooting`
Smoke tests	`opp-repl-smoke-tests`
Fingerprint tests	`opp-repl-fingerprint-tests`
Statistical tests	`opp-repl-statistical-tests`
Speed tests	`opp-repl-speed-tests`
Chart tests	`opp-repl-chart-tests`
Module-image tests	`opp-repl-module-image-tests`
Sanitizer tests	`opp-repl-sanitizer-tests`
Feature + release tests	`opp-repl-feature-and-release-tests`
Comparing simulations	`opp-repl-comparing-simulations`
Parameter optimization	`opp-repl-parameter-optimization`
Coverage reports	`opp-repl-coverage-reports`
Profiling (perf/Hotspot)	`opp-repl-profiling`
Overlay builds	`opp-repl-overlay-builds`
SSH / Dask cluster runs	`opp-repl-ssh-cluster`
GitHub Actions dispatch	`opp-repl-github-actions`
opp_env integration	`opp-repl-opp-env-integration`
MCP server for AI agents	`opp-repl-mcp-server`
Shared human+AI REPL	`opp-repl-shared-terminal`
Sandboxed execution	`opp-repl-sandbox`
End-to-end AI recipes	`opp-repl-ai-workflows`

Upstream API touchpoints worth knowing

Three high-value APIs on current opp_repl (Apr 2026+) that the skills route you to:

create_project(name, path=..., namespace=False) — scaffold a new project in one call. See opp-repl-project-scaffolding.
build_project(simulation_project=p) — builds. Auto-generates the Makefile from .oppbuildspec on the first call.
r.get_scalars() / r.get_vectors() / r.get_histograms() — read results into a pandas DataFrame directly from a SimulationTaskResult or MultipleSimulationTaskResults.

Minimum viable skill set for a first session

If you're not sure what to load, start with this core set — it covers 80% of real-world opp_repl usage:

opp-repl-concepts                 (object model)
opp-repl-opp-files                (project descriptors)
opp-repl-running-simulations      (the main entry point)
opp-repl-tasks-and-results        (inspecting what happened)
opp-repl-result-analysis          (reading .sca files)
opp-repl-troubleshooting          (decoding errors)

Add one or two test-type skills when testing, or opp-repl-mcp-server when connecting an agent.

When NOT to use opp_repl

One-off manual runs — plain opp_run or the OMNeT++ IDE is simpler.
Pure analysis of existing .sca files with no rerun — just use opp_scavetool directly.
Simulations that require GUI interaction during the run — opp_repl is headless.

Authoritative upstream references

Repo: https://github.com/omnetpp/opp_repl
Docs: https://github.com/omnetpp/opp_repl/tree/main/doc
Each sibling skill cites the specific upstream doc file it distils.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most mcp tooling skills give in ~2.0k tokens

Counted across 638 of the 750 authors here whose files we hold, read 2026-08-07

Create ten complex or independent read-only evaluation questionsin 69 of 638, across 15 files
Test servers using MCP Inspectorin 61 of 638, across 19 files
Provide actionable error messages with specific next stepsin 54 of 638, across 12 files
Prioritize comprehensive API coverage over specific workflows or workflow toolsin 54 of 638, across 12 files
Use TypeScript and Streamable HTTP for remote servers or clientsin 54 of 638, across 8 files
Define structured output schemas where possiblein 50 of 638, across 8 files
Use Zod or Pydantic for input schemasin 47 of 638, across 5 files
Fetch MCP specification pages with markdown suffixin 46 of 638, across 4 files
Load framework documentation using WebFetchin 45 of 638, across 3 files
Verify each evaluation answer independentlyin 45 of 638, across 3 files
Implement API client with authentication and paginationin 45 of 638, across 3 files
Define input schemas with validationin 27 of 638, across 9 files

Said here and by no other author read

load this skill first for opp_repl tasks
load exactly the specified sibling skills for a task
load the core skill set for a first session
read the decision tree to triage tasks
load troubleshooting skill first for failures
use opp_scavetool for pure existing sca analysis

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.