APEX
10-Agent Senior Engineering Team for Any CLI Coding Agent
A zero-dependency orchestrator that routes your requests to 10 specialist AI agents — each with domain expertise, purpose-built tools, and the ability to dynamically call peers mid-task. Works with 16+ coding agents out of the box.
Quick Start • Agents • Installation • Skills • MCP Servers • Commands • Composio • Mirage VFS • License
Quick Start
# Install globally via npm
npm install -g @asno-dev/apex
# Run universal installer — detects all your coding agents
npx @asno-dev/apexThat's it. Now talk to your coding agent:
@arch refactor this → Max compresses your code
@ui build a login form → Zara paints a WCAG AA form
@debug fix this error → Kai runs 5-step debug protocol
@perf this is slow → Rex profiles & optimizes
@sec review auth code → Vex scans for OWASP Top 10
@infra dockerize this → Io outputs production-grade config
@nova any ideas? → Nova proposes non-obvious angles
@reed best caching strategy → Dr. Reed compares options with evidence
@review check this code → Rila gives structured PR review
@flex what's the MVP? → Flex scores Value×Cost and cuts scope
Built-in Commands
apex-docs → Create Word/Excel/PPT documents
apex-excel
apex-ppt
apex-composio-setup → Connect 1000+ external tools (Gmail, GitHub, Slack, etc.)
apex-composio-status → Show connected tools status
apex-composio-sync → Force sync from Composio backend
apex-mirage ls /s3/ → Virtual filesystem across 50+ backends
/apex team → Switch to team mode (default)
/apex select a,b → Activate only specific agents
/apex status → Show current mode and active agents
Agents
APEX ships with 10 specialist agents, each with a unique persona, domain expertise, and set of purpose-built tools.
| Badge | Tag | Name | Role | Specialty |
|---|---|---|---|---|
[Arch] |
@arch |
Max | Architect | System design, refactoring, structure. Compresses 50→1 line. |
[UI] |
@ui |
Zara | UI/UX Designer | Mood-first, anti-slop. shadcn/ui + Tailwind. WCAG AA. 10 palettes. |
[Dbg] |
@debug |
Kai | Debugger | 5-step protocol: reproduce → isolate → hypothesize → fix → prevent. |
[Perf] |
@perf |
Rex | Performance Engineer | Profile first. Algorithm → DB → bundle → render. Baseline → optimize → measure. |
[Sec] |
@sec |
Vex | Security Engineer | OWASP Top 10. CRITICAL/HIGH/MEDIUM. Every input is malicious. |
[Inf] |
@infra |
Io | Infrastructure | Docker/k8s/CI-CD. Multi-stage builds. Non-root. Rollback always. |
[Nov] |
@nova |
Nova | Creative | Non-obvious angles. Lib + npm + POC + downside analysis. |
[Res] |
@reed |
Dr. Reed | Researcher | Evidence-based. ≥2 options with O(?) complexity. No opinions without data. |
[Rev] |
@review |
Rila | Code Reviewer | Blocking → Suggestions → Praise. Specific praise always. |
[Fnd] |
@flex |
Flex | Founder / PM | Value(1-3) × Cost(1-3). Ships 60%, defers 30%, kills 10%. |
Automatic Routing
The orchestrator analyzes your request and routes to the best agent automatically:
| Your Request | Routed To |
|---|---|
| Code is long/complex/messy | @arch |
| Build UI / component / page | @ui |
| Error / bug / crash / undefined | @debug |
| Slow / memory / performance | @perf |
| Auth / input / secrets / vuln | @sec |
| Deploy / Docker / CI / k8s | @infra |
| New idea / library / creative | @nova |
| Best way / which / research | @reed |
| Review / PR / merge / quality | @review |
| Scope / MVP / what to build | @flex |
Task States
Every agent reports progress with a single icon:
🧠 Thinking → 🔍 Exploring → ⚡ Working → 🔧 Fixing → ✅ Verifying → ✨ Complete
Modes
APEX supports three operating modes:
Team Mode (Default)
The orchestrator routes your request to the best agent. That agent works and dynamically calls peers when needed. Zero pre-loading.
> refactor and optimize the API layer
# Orchestrator routes → @arch → @arch calls @perf when spotting N+1 queries
Direct Mode
Address a specific agent directly. That agent becomes the lead with full authority.
@debug fix the authentication timeout
# Kai works directly, can still call @sec if a vulnerability is found
Select Mode
Activate only specific agents for a focused session.
/apex select kai,rex # Only debugger and performance active
/apex select arch,ui,infra # Architecture + UI + Infrastructure
Mode Commands
| Command | Description |
|---|---|
/apex team |
Switch to team mode (default) |
/apex select a,b |
Activate only specific agents |
/apex off |
Disable APEX |
/apex status |
Show current mode and active agents |
/apex help |
Show help guide |
Installation
APEX supports 16+ coding agents with dedicated adapters for each.
Universal Installer (Recommended)
# Auto-detect installed agents and configure all of them
npx @asno-dev/apex
# Or install globally
npm install -g @asno-dev/apexThe CLI auto-detects which coding agents are installed on your system and copies the right configuration files for each one. It handles all the format differences — plugin.json for Claude Code, .mdc for Cursor, plugin.json + toml for Codex, extension.json + toml for Gemini, and so on.
Shell Installers
# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/asno-dev/apex/main/install.sh | sh
# Windows (PowerShell 7+)
.\install.ps1Per-Agent Details
| Agent | Config File(s) | Notes |
|---|---|---|
| Claude Code | ~/.claude/plugin.json + ~/.claude/agents/* + ~/.claude/commands/* |
10 subagents, 8 commands, 3 MCP servers |
| Codex CLI | ~/.codex/plugin.json + ~/.codex/agents/*.toml + ~/.codex/mcp.toml |
10 subagents, 8 commands, 3 MCP servers |
| Cursor | .cursor/rules/apex.mdc + .cursor/agents/*.mdc + .cursor/commands/*.md |
10 agents, 8 commands, MCP |
| Gemini CLI | ~/.gemini/extension.json + ~/.gemini/agents/* + ~/.gemini/commands/*.toml |
10 agents, 8 commands, 3 MCP servers |
| Windsurf | .windsurf/rules/apex.md + .windsurf/agents/* + .windsurf/workflows/* |
10 agents, 8 workflows |
| Cline / Kilo Code | .clinerules (project root) |
Comprehensive rules file |
| GitHub Copilot | .github/copilot-instructions.md |
Full APEX instructions |
| GitHub Copilot CLI | ~/.config/gh-copilot/config.yml (via CLI plugin) |
Subcommands |
| Devin | ~/.devin/plugin.yaml + ~/.devin/mcp.json + ~/.devin/agents/* |
10 agents, MCP |
| Hermes | ~/.hermes/apex-features.yaml + ~/.hermes/plugin.yaml |
Feature flags |
| OpenCode | opencode.json + .opencode/* |
Native subagent support |
| Pi Agent | ~/.pi/extension.json |
MCP servers |
| Antigravity | ~/.agy/antigravity-extension.json |
Extension + MCP |
| Kiro | ~/.kiro/steering/apex.md |
Steering instructions |
| Swival | ~/.swival/apex.md (via skills add) |
Skill rules |
| OpenClaw | ~/.clawhub/apex.md + ~/.clawhub/package.json |
Package + MCP |
| CodeWhale | AGENTS.md (project root) |
Auto-reads |
Commands
All APEX commands use the apex- prefix for clean namespacing — no collisions with built-in agent commands.
| Command | Description |
|---|---|
apex-docs |
Create/edit Word documents via OfficeCLI |
apex-excel |
Create/edit Excel spreadsheets via OfficeCLI |
apex-ppt |
Create PowerPoint presentations via OfficeCLI |
apex-composio-setup |
Connect external tools — paste API key, get OAuth link |
apex-composio-status |
Show connected tools and API key status |
apex-composio-sync |
Force sync from Composio backend |
apex-mirage <command> |
Execute commands across mounted virtual filesystem backends |
/apex team|select|off|status|help |
APEX mode control |
Skills
APEX includes 25 pre-built skills — composable, multi-step workflows that any agent can invoke.
Agent Skills (10)
Each agent has its own core skill defining its persona, behavior, and domain rules:
| Skill | Agent | Description |
|---|---|---|
apex-arch |
Max | System architecture, refactoring heuristics, blast radius mapping |
apex-ui |
Zara | UI/UX design system, palette selection, component painting |
apex-debug |
Kai | 5-step debug protocol, stack walking, guard injection |
apex-perf |
Rex | Profiling, baseline capture, Big-O analysis |
apex-sec |
Vex | OWASP scanning, secret detection, input tracing |
apex-infra |
Io | Docker linting, k8s validation, CI/CD pipeline checks |
apex-nova |
Nova | POC generation, library compass, trend analysis |
apex-reed |
Dr. Reed | Evidence search, tradeoff matrices, complexity calculation |
apex-review |
Rila | Diff categorization, anti-pattern detection, quality gates |
apex-flex |
Flex | Value-cost scoring, MVP cutting, effort estimation |
Workflow Skills (15)
Pre-composed multi-agent workflows for common engineering tasks:
| Skill | Agents Used | Description |
|---|---|---|
refactor |
@arch | Code compression and structural refactoring |
ui-generate |
@ui | Generate UI components from descriptions |
debug-protocol |
@debug | Full 5-step debugging workflow |
perf-audit |
@perf | End-to-end performance audit |
sec-review |
@sec | Security review with OWASP scoring |
pr-review |
@review | Structured pull request review |
system-design |
@arch, @reed | Architecture design with research backing |
research |
@reed | Evidence-based technology research |
mvp-cut |
@flex | MVP scoping and feature prioritization |
full-stack-scaffold |
@arch → @ui → @infra | Complete app scaffolding pipeline |
test-harness |
@debug | Test suite generation |
bug-patch |
@debug → @review | Bug fix with review verification |
requirements-to-spec |
@flex → @arch | Requirements analysis to technical spec |
deploy-pipeline |
@infra | CI/CD pipeline generation |
MCP Servers
APEX provides 3 MCP (Model Context Protocol) servers with 56 purpose-built tools across all 10 agents, plus 6 virtual filesystem tools and 1000+ external tool bridges.
1. apex-hands — 56 Agent Domain Tools
The primary MCP server. 56 tools organized by agent domain:
| Agent | Tools | Description |
|---|---|---|
@arch |
blast_radius dep_graph complexity extract_refactor compose_check module_boundary |
Architecture analysis |
@ui |
contrast palette_extract a11y_audit responsive_test component_search |
UI/UX tooling |
@debug |
reproduce stack_walk log_mine bisect_run guard_inject var_watch |
Debugging instruments |
@perf |
profile memory_profile baseline_capture measure bundle_analyze big_o |
Performance profiling |
@sec |
vuln_scan secret_find input_trace auth_map owasp_score dependency_audit |
Security scanning |
@infra |
docker_lint k8s_validate ci_check deploy_dry rollback_plan health_check |
Infrastructure ops |
@nova |
poc_gen lib_compass alt_angle trend_sniff downside_check approach_matrix |
Creative exploration |
@reed |
compare complexity_calc evidence_search tradeoff_matrix recommend |
Research tools |
@review |
diff_cat anti_pattern quality_gate praise_find review_card |
Code review |
@flex |
value_cost mvp_cut risk_matrix roadmap effort_estimate |
Product management |
MCP Config:
{
"mcpServers": {
"apex-hands": {
"type": "local",
"command": ["node", "src/hands-server.mjs"]
}
}
}2. mirage-vfs — Virtual Filesystem (50+ Backends)
Unified filesystem across S3, GDrive, Slack, Redis, Postgres, and more.
MCP Config:
{
"mcpServers": {
"mirage-vfs": {
"type": "local",
"command": ["node", "src/mirage-server.mjs"]
}
}
}Usage:
apex-mirage ls /s3/ — List files in S3 bucket
apex-mirage cp /gdrive/report.pdf /data/ — Copy from Google Drive
apex-mirage grep -r error /s3/logs/ — Search across backends
apex-mirage cat /slack/channel/messages — Read Slack messages
Setup: pip install mirage-ai && npm install -g @struktoai/mirage-cli
3. apex-composio — 1000+ External Tool Bridge
Bridges Gmail, GitHub, Slack, Jira, Notion, and 1000+ more tools.
MCP Config:
{
"mcpServers": {
"apex-composio": {
"type": "local",
"command": ["node", "src/composio-server.mjs"]
}
}
}All 3 MCP Servers in One Config
The root .mcp.json includes all three servers:
{
"mcpServers": {
"apex-hands": { "command": ["node", "src/hands-server.mjs"] },
"mirage-vfs": { "command": ["node", "src/mirage-server.mjs"] },
"apex-composio": { "command": ["node", "src/composio-server.mjs"] }
}
}Composio Integration
Connect 1000+ external tools (Gmail, GitHub, Slack, Google Drive, Jira, Linear, Notion, and more) through Composio.
Setup
# Interactive setup wizard
apex-composio-setupThe wizard will:
- Ask for your Composio API key (get one at composio.dev)
- Let you select which tools to connect (gmail, github, slack, jira, etc.)
- Open an OAuth link in your browser for authorization
- Save the connection config to
.composio-config.json
Status & Sync
apex-composio-status # Show all connected tools and their status
apex-composio-sync # Force refresh tool definitions from backendUsage
After connecting a tool, invoke it with @toolName:
@gmail send an email to the team about the release
@github create a PR from feature branch
@slack post build status to #deployments
@jira create a ticket for this bug
@notion update the architecture doc
Configuration
Stored at .composio-config.json (auto-added to .gitignore / .npmignore):
{
"apiKey": "ak_your_api_key",
"userId": "your_user_id",
"connections": [
{ "tool": "gmail", "label": "Gmail", "status": "ACTIVE", "authType": "OAUTH2" }
]
}Agent Chains
Agents can be chained for complex workflows:
Sequential Chains
Full app: @arch → @ui → @infra # Design → Paint → Deploy
Bug patch: @debug → @review # Fix → Verify
Spec to code: @flex → @arch → @ui # Scope → Design → Build
Parallel Chains
System design: @arch ∥ @reed # Architecture + Research in parallel
Dynamic Peer Calling
Any agent can call any peer mid-task:
@perffinds SQL injection → calls@sec@uineeds a backend API → calls@infra@debugfinds performance issue → calls@perf@reviewfinds security concern → calls@sec
Programmatic API
Use APEX as a Node.js library:
const apex = require('@asno-dev/apex');
console.log(apex.version); // "3.0.2"
console.log(apex.agents); // ['arch', 'ui', 'debug', 'perf', 'sec', 'infra', 'nova', 'reed', 'review', 'flex']
console.log(apex.listAdapters()); // ['claude-code', 'cursor', 'opencode', 'cline', 'copilot', ...]
const archSkill = apex.getSkill('arch');
const cursorAdapter = apex.getAdapter('cursor');Design System (Zara)
Zara's UI design system includes:
- 10 curated palettes: Trust, Energy, Authority, Clarity, Warmth, Midnight, Forest, Ocean, Aurora, Minimal
- CSS variable tokens for all colors as
:rootvariables - 2 fonts max per project
- shadcn/ui component library + Tailwind CSS scale
- WCAG AA compliance (4.5:1 contrast ratio minimum)
- Mobile-first responsive design
- 200ms max transition duration
- Skeleton loaders for async content
- Semantic HTML throughout
Anti-Slop Rules
- No decorative elements without purpose
- No inline styles
- No hardcoded hex colors
- No lorem ipsum in production
- No gradients without justification
- No ALL CAPS text
Requirements
- Node.js ≥ 18.0.0
- At least one supported coding agent installed
Zero Dependencies
APEX has zero production dependencies. It uses only Node.js built-in modules:
- Instant install
- No supply chain risk
- Tiny package size
Contributing
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
Guidelines
- Follow the Core Laws in your contributions
- Keep adapters in sync — run
npm testbefore submitting - Add skills with proper YAML frontmatter
- Document new tools in the appropriate
hands/*.mjsmodule
Development
git clone https://github.com/asno-dev/apex.git
cd apex
npm testProject Structure
| Directory | Purpose |
|---|---|
adapters/ |
Per-agent config files (16 agents) |
agents/ |
Canonical agent definitions (10 agents) |
commands/ |
Canonical command definitions (8 commands) |
skills/ |
SKILL.md files for orchestrator and agents |
src/ |
MCP servers and tool implementations |
src/hands/ |
Individual tool modules per agent |
hooks/ |
Lifecycle hooks for session management |
scripts/ |
Build and validation scripts |
docs/ |
Extended design documentation |
bin/ |
CLI entry point |
License
Built with by asno-dev