Maintainer of this skill?
Automated security scan — not a certification.
by cluster1900
Apexo: An open-source, general-purpose coding agent for terminal (TUI), web, and desktop with autonomous goals, workflows, persistent memory, command sandboxing, and first-class support for Gemini, Grok, and OpenAI-compatible models.
# Add to your Claude Code skills
git clone https://github.com/cluster1900/apexo-codeLast scanned: 10/31/2026
{
"issues": [
{
"file": "README.md",
"line": 87,
"type": "remote-install",
"message": "Install command (remote install script piped to a shell — review the source before running): \"curl -fsSL https://raw.githubusercontent.com/cluster1900/apexo-code/main/install\"",
"severity": "low"
}
],
"status": "PASSED",
"scannedAt": "2026-10-31T10:12:15.904Z",
"npmAuditRan": false,
"pipAuditRan": true,
"promptInjectionRan": true
}See how apexo-code compares with popular alternatives.
apexo-code is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by cluster1900. Apexo: An open-source, general-purpose coding agent for terminal (TUI), web, and desktop with autonomous goals, workflows, persistent memory, command sandboxing, and first-class support for Gemini, Grok, and OpenAI-compatible models. It has 223 GitHub stars.
Yes. apexo-code passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/cluster1900/apexo-code" and add it to your Claude Code skills directory (see the Installation section above).
apexo-code is primarily written in TypeScript. It is open-source under cluster1900 on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh apexo-code against similar tools.
No comments yet. Be the first to share your thoughts!
From $49 · 30 days · terms & refunds
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
Automated security scan — not a certification.
Apexo is a general-purpose coding agent with a unified terminal UI (TUI), web UI, and Electron desktop app powered by a single local agent runtime.
Rather than being tied to a single model vendor or acting as a thin wrapper, Apexo provides its own deeply engineered agent architecture — including a durable session runtime, autonomous goals (/goal), multi-session workflows (/workflow), background task & subagent orchestration (/tasks), persistent project memory (/memory), native command sandboxing (/sandbox), lifecycle hooks (/hooks), and side-channel queries (/btw) — while offering first-class support across major model ecosystems:
agy) Google login directly without re-authorizing, or connect via GOOGLE_GENERATIVE_AI_API_KEY. Full support for Gemini models, thinking/reasoning summaries, multimodal attachments, and large-context coding workflows. See the Antigravity setup guide.XAI_API_KEY. All current Grok models (grok-4.7, grok-4.6, grok-4.20, grok-build-0.1) and Grok Build-style agent capabilities work natively.<think> reasoning extraction, incremental streaming, and tool execution.Apexo supports local OAuth / native CLI login for Google Antigravity (agy) and xAI Grok, plus direct API-key and custom endpoint connections for all providers (Google Gemini, xAI, OpenAI, Anthropic, MiniMax, and any OpenAI-compatible service). Authentication is handled entirely on your machine; model requests go straight to the provider you choose.
/goal): Define a target objective and verification criteria; the agent iterates until the goal is verified, paused, blocked, or out of budget./workflow): Run structured multi-step workflows in dedicated child sessions while keeping your main conversation responsive./tasks): Manage background shell commands, scheduled recurring jobs (/loop), subagents, and coordinator sessions in one place./memory, /remember, /flush, /dream): Review, curate, and consolidate durable project and user memory across sessions./sandbox): Enforce OS-level sandboxing (seatbelt on macOS, bubblewrap on Linux) for native shell commands (workspace or read-only profiles)./hooks): Inspect and govern repository and user hooks with explicit per-repository trust controls./btw): Ask a quick side question (even mid-prompt) in a modal without polluting or interrupting the active session turn.npm install -g apexo-code
apexo --version
The npm package installs a prebuilt CLI for macOS, Linux (glibc or musl), or Windows on x64 and arm64. Desktop installers and standalone CLI archives are available from GitHub Releases.
See the detailed installation guide for platform and CPU selection, npm-free installation, Windows PowerShell commands, checksums, PATH, Desktop installation, updates, and removal.
To build Apexo from source, install Bun 1.4.4 (the repo pins bun@1.4.4) and git.
git clone https://github.com/cluster1900/apexo-code.git apexo && cd apexo
bun install
# Build a single native binary for this machine (the web UI is embedded)
bun run --cwd packages/apexo build --single
# The binary lands in packages/apexo/dist/apexo-<os>-<arch>/bin/apexo
mkdir -p ~/.local/bin
install -m755 packages/apexo/dist/apexo-*/bin/apexo ~/.local/bin/apexo
install -m755 packages/apexo/dist/apexo-*/bin/apexo-workflow-script ~/.local/bin/apexo-workflow-script
export PATH="$HOME/.local/bin:$PATH"
apexo --version
Pushing a vX.Y.Z tag runs .github/workflows/release.yml, which builds and smoke-tests the CLI npm packages, packages Desktop for macOS, Linux, and Windows,
and attaches the archives to a GitHub Release.
See Releasing Apexo for publishing setup and the release checklist.
Once a release exists, the install script can fetch it instead (--version X.Y.Z pins a release):
curl -fsSL https://raw.githubusercontent.com/cluster1900/apexo-code/main/install | bash
To run from source without building:
bun dev # same as: bun run --cwd packages/apexo src/index.ts
bun dev --help
apexo # start the terminal UI in the current directory
apexo auth login # sign in to Google Antigravity, xAI Grok (OAuth), or configure an API key
apexo auth list # show configured credentials
apexo models # list available models across all configured providers
apexo run "explain this repo" # one-shot, non-interactive CLI execution
apexo web # start the local server and open the web UI
apexo serve # headless local server (for desktop app, editors, scripts)
apexo acp # run Apexo as an Agent Client Protocol (ACP) server for editors like Zed
In the TUI and Desktop composer, /connect opens the provider dialog, /models switches models, /themes switches
themes, and ctrl+p (cmd+k / cmd+p on macOS) opens the command palette. Start a prompt with ! to run a shell command.
| Provider | How to connect | Environment variable |
|---|---|---|
| Google Gemini | API key (AI Studio / Vertex-compatible) | GOOGLE_GENERATIVE_AI_API_KEY |
| Antigravity (Google / Gemini) | Select Antigravity; reuse local agy login or complete agy's Google sign-in once |
— |
| xAI Grok | apexo auth login -> xAI -> SuperGrok Subscription (device-code OAuth), or an API key |
XAI_API_KEY |
| OpenAI | API key | OPENAI_API_KEY |
| Anthropic (Claude) | API key | ANTHROPIC_API_KEY |
| Custom / OpenAI-Compatible | Configure baseURL + apiKey in apexo.jsonc (MiniMax, DeepSeek, Ollama, vLLM, etc.) |
Custom {env:...} |
The Electron desktop app lives in packages/desktop:
bun run --cwd packages/desktop dev # development build
bun run --cwd packages/desktop build # production bundle (packaging uses electron-builder)
The desktop app has no auto-updater; rebuild it (or install a newer release) to update.
Apexo reads JSON/JSONC config files:
| What | Location |
|---|---|
| Global config | ~/.config/apexo/apexo.jsonc (apexo.json also works) |
| Data / state / cache | ~/.local/share/apexo, ~/.local/state/apexo, ~/.cache/apexo |
| Project config | apexo.json / apexo.jsonc in the project (searched up to the git root) |
| Project directory | .apexo/ (agents, commands, plugins, themes, tools) |
| Environment variables | APEXO_* (for example APEXO_CONFIG, APEXO_CONFIG_CONTENT) |
Example apexo.json:
{
"$schema": "https://raw.githubusercontent.com/cluster1900/apexo-code/main/schemas/config.json",
"model": "google/gemini-2.5-pro",
"theme": "grove",
}
You can configure any self-hosted or third-party OpenAI-compatible endpoint (such as MiniMax, DeepSeek, Qwen, vLLM, Ollama, or LM Studio) in ~/.config/apexo/apexo.jsonc or your project's apexo.jsonc:
{
"$schema": "https://raw.githubusercontent.com/cluster1900/apexo-code/main/schemas/config.json",