Evolver CLI

Inspect, generate, and run evolution workflows in your terminal

Run the documented Evolver self-evolution engine, inspect its output, and keep every local or network effect inside an explicit boundary.

$ evolver

Evolver CLI · local evolution workflow

project: ./your-test-project

Describe the goal, allowed scope, and validation…

Work in a local project

Run Evolver against a bounded repository or agent workflow and keep relevant context close to the task.

Stay in control

Choose strategy, loops, file boundaries, validation, publishing, and credit behavior before the run.

Export reusable history

Review Genes, Capsules, Events, Skills, and portable .gepx archives without treating the CLI as an EvoX edition.

Quickstart

Get started with Evolver CLI

Install the package, verify the command, and begin in a non-production project.

Install Evolver CLI
Check the installed package metadata for Node requirements. The verified 2.0.28 package requires Node.js ^22.13.0 or >=23.4.0, plus Git.
```
npm install -g @evomap/evolver
```
Verify the installation
Confirm the command is available before opening a project or enabling network behavior.
```
evolver --help
```
Open a bounded project
Choose a test repository, document the allowed files and commands, and remove production credentials.
```
cd <your-test-project>
```
Run your first task
Run a single evolution cycle that scans signals, selects a matching Gene or Capsule, emits a GEP prompt, and records an EvolutionEvent.
```
evolver
```

See what Evolver CLI can do

Complete an evolution cycle in the terminal

Scan local logs and signals, select reusable experience, and emit a protocol-bound prompt for the agent application that consumes it.

$ evolver

inspect input · retain records · define boundaries

Use Genes, Capsules, and skills

Reuse validated strategies and outcomes, then turn stable recurring approaches into portable assets.

$ evolver fetch --skill <id>

inspect input · retain records · define boundaries

Review before continuing

Use review mode and inspect the generated prompt, events, and proposed effects before a host applies changes.

$ evolver --review

inspect input · retain records · define boundaries

Build terminal workflows around Evolver

Review project signalsScan the local memory directory for logs, failures, and other signals relevant to the next cycle.

Bring in local contextRun inside a Git repository so the workflow can use bounded project context and preserve an audit trail.

Generate a GEP promptSelect a matching Gene or Capsule and emit a structured prompt for the connected agent application.

Inspect events and assetsReview EvolutionEvents and resulting local assets before deciding whether to reuse or publish them.

Connect to EvoMap Hub when neededKeep local mode offline, or explicitly configure identity and network access for sharing and collaboration.

Integrate through hooks or MCPConnect supported agent applications and tools while keeping credentials and external effects explicit.

Export a portable evolution archive

Evolver CLI can package Evolver/GEP evolution history into a Portable Evolution Archive. Keep node credentials out of the archive and verify the resulting file before transfer.

evolver sync --export backup.gepx

Use Evolver CLI in automation

For CI or scheduled execution, pin the tool version, inject identity securely, set hard runtime and change limits, keep publishing and spending explicit, and preserve logs and artifacts for review.

Use Evolver CLI when…

You work in a local Git projectRun auditable evolution cycles close to project logs and signals.

You need reusable experienceRecall or produce Genes, Capsules, Skills, Events, and portable archives.

You want human reviewPause in review mode before the connected agent application continues with the proposed workflow.

You need repeatable automationUse explicit loop, hook, MCP, or CI boundaries and retain the resulting audit trail.

Next steps

Configure EvolverReview identity, strategy, loops, publishing, credits, and platform paths.Review permissionsDefine local, command, network, publishing, and approval boundaries.Export .gepxBack up or migrate verified Evolver/GEP evolution history.

Get started with Evolver CLI

Install the package, verify the command, and begin in a non-production project.

Install Evolver CLI

Check the installed package metadata for Node requirements. The verified 2.0.28 package requires Node.js ^22.13.0 or >=23.4.0, plus Git.

npm install -g @evomap/evolver

Verify the installation

Confirm the command is available before opening a project or enabling network behavior.

evolver --help

Open a bounded project

Choose a test repository, document the allowed files and commands, and remove production credentials.

cd <your-test-project>

Run your first task

Run a single evolution cycle that scans signals, selects a matching Gene or Capsule, emits a GEP prompt, and records an EvolutionEvent.

evolver

See what Evolver CLI can do

Complete an evolution cycle in the terminal

Scan local logs and signals, select reusable experience, and emit a protocol-bound prompt for the agent application that consumes it.

$ evolver

inspect input · retain records · define boundaries

Use Genes, Capsules, and skills

Reuse validated strategies and outcomes, then turn stable recurring approaches into portable assets.

$ evolver fetch --skill <id>

inspect input · retain records · define boundaries

Review before continuing

Use review mode and inspect the generated prompt, events, and proposed effects before a host applies changes.

$ evolver --review

inspect input · retain records · define boundaries

Build terminal workflows around Evolver

Review project signalsScan the local memory directory for logs, failures, and other signals relevant to the next cycle.

Bring in local contextRun inside a Git repository so the workflow can use bounded project context and preserve an audit trail.

Generate a GEP promptSelect a matching Gene or Capsule and emit a structured prompt for the connected agent application.

Inspect events and assetsReview EvolutionEvents and resulting local assets before deciding whether to reuse or publish them.

Connect to EvoMap Hub when neededKeep local mode offline, or explicitly configure identity and network access for sharing and collaboration.

Integrate through hooks or MCPConnect supported agent applications and tools while keeping credentials and external effects explicit.

Use Evolver CLI when…

You work in a local Git projectRun auditable evolution cycles close to project logs and signals.

You need reusable experienceRecall or produce Genes, Capsules, Skills, Events, and portable archives.

You want human reviewPause in review mode before the connected agent application continues with the proposed workflow.

You need repeatable automationUse explicit loop, hook, MCP, or CI boundaries and retain the resulting audit trail.