4cli-unified

name	4cli-unified
description	Unified meta-CLI wrapper for OpenCode + Claude Code + Codex + Gemini CLI — one command for four agents with 9Router tier routing and unified credential vault. Use when use when managing multiple ai coding clis, setting up unified routing across opencode/claude/codex/gemini, or building credential vaults Triggers on keywords: cli, unified, opencode, claude, codex, gemini, owl.
allowed-tools	Read, Write, Bash, Edit, Grep, Glob
user-invocable	true
mcp-servers	["filesystem","github","docker"]

Unified Meta-CLI

context

Unified meta-CLI wrapper for OpenCode + Claude Code + Codex CLI + Gemini CLI. Provides consistent interface across AI coding assistants with shared config and history. This skill provides infrastructure-as-code patterns and operational procedures that ensure reliability, security, and observability of critical infrastructure components.

instructions

Step 1: Assess Infrastructure Requirements

1. Identify the infrastructure need — routing, proxying, credential management, or CLI orchestration
2. Determine deployment target — local, cloud, or hybrid
3. Map dependencies — what services, APIs, or systems must be accessible
4. Establish SLOs — uptime, latency, and error rate targets

Step 2: Design Infrastructure Layer

1. Architecture — Define the component topology and data flow
2. Configuration — Set up environment-specific configuration with secrets management
3. Health checks — Implement liveness, readiness, and startup probes
4. Failover — Design redundancy with automatic recovery procedures

Step 3: Generate Infrastructure Specification

## Infrastructure Specification

### Architecture
Component diagram and data flow description

### Configuration
- Environment variables: list with descriptions
- Secrets: list with rotation policy
- Resource limits: CPU, memory, disk

### Health and Monitoring
- Liveness: health check endpoint and interval
- Readiness: dependency connectivity check
- Metrics: key metrics to track
- Alerts: threshold-based alerting rules

### Failover Procedures
- Primary failure: automatic recovery steps
- Cascading failure: circuit breaker behavior
- Full outage: manual recovery runbook

Step 4: Operational Readiness

• Configuration is externalized and documented
• Secrets are encrypted at rest and in transit
• Health checks cover all critical dependencies
• Failover is tested and automated
• Monitoring and alerting are configured
• Runbook exists for common failure scenarios

constraints

• NEVER store secrets in code or configuration files — use a vault.
• NEVER deploy without health checks and monitoring.
• NEVER assume network reliability — design for failure at every layer.
• ALWAYS test failover procedures before relying on them.
• ALWAYS maintain audit logs for all credential access.

examples

Example: Infrastructure Setup

Input: "Use when managing multiple AI coding CLIs, setting up unified routing across OpenCode/Claude/Codex/Gemini, or building credential vaults"

Output:

## Infrastructure Specification

### Architecture
Multi-layer 4cli unified with automatic failover between primary and backup endpoints.

### Configuration
- PRIMARY_ENDPOINT: configured via environment
- FALLBACK_ENDPOINT: configured via environment
- MAX_RETRIES: 3
- TIMEOUT_MS: 5000

### Health and Monitoring
- Liveness: GET /health every 30s
- Readiness: Dependency connectivity check
- Metrics: request_rate, error_rate, latency_p99
- Alerts: error_rate exceeding threshold for 2 minutes

### Failover Procedures
- Primary failure: Auto-switch to fallback within 5s
- Both down: Queue requests, retry every 30s
- Recovery: Auto-switch back to primary when healthy