cli-opencode

Dispatch external AI runtimes into full OpenCode sessions with plugins, skills, and MCPs.

31|3|Updated Dec 23, 2025
One-click install
npx skills add https://github.com/MichelKerkmeester/opencode--spec-kit-skilled-agent-orchestration --skill cli-opencode
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cli-opencode
Source: https://github.com/MichelKerkmeester/opencode--spec-kit-skilled-agent-orchestration/tree/main/.opencode/skill/cli-opencode
Command: npx skills add https://github.com/MichelKerkmeester/opencode--spec-kit-skilled-agent-orchestration --skill cli-opencode

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) and references (resource) components.

What problem does it solve?

The OpenCode CLI orchestrator enables external AI runtimes to dispatch into a full OpenCode session, loading the project's plugins, skills, MCPs, and Spec Kit Memory runtime, while supporting parallel detached sessions and cross-AI handback.

Core Features & Use Cases

  • Full plugin/skill/MCP runtime dispatch for external AI runtimes, enabling access to the project's specifications, memory tools, and code graph.
  • In-OpenCode parallel detached sessions for ablation suites, worker farms, and long-running experiments, with separate session state and optional share URLs.
  • Cross-AI orchestration handback to non-Anthropic CLIs (Codex, Copilot, Gemini) enabling spec-kit workflows that require OpenCode runtime context.
  • Self-invocation guard that prevents circular delegation when OpenCode is already active, with a defined exception for explicit parallel detached requests.
  • Cross-repo and directory routing via --dir to enable project-to-project runtime handoffs without leaving the host environment.

Quick Start

Invoke the OpenCode CLI from an external AI with the canonical dispatch shape to start a fresh session and load the runtime context.

Frequently Asked Questions about cli-opencode

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I dispatch tasks from external AI runtimes like Codex or Gemini into a full project session?

Cross-AI dispatch works by invoking the OpenCode CLI orchestrator from external runtimes, loading the project's plugins, skills, MCPs, and Spec Kit Memory into a single session while preserving memory and code graph context.

Can I run parallel detached sessions for long-running experiments without losing session state?

Parallel detached sessions are supported for ablation suites and worker farms, maintaining separate session state with optional share URLs, while a self-invocation guard prevents circular delegation unless explicitly requested.

What is cross-repo dispatch and how does --dir routing work for project-to-project handoffs?

Cross-repo dispatch via the --dir flag enables project-to-project runtime handoffs, allowing external AI runtimes to route tasks to different directories without leaving the host environment.

How does the self-invocation guard handle circular delegation when OpenCode is already active?

The self-invocation guard prevents circular delegation when OpenCode is already active, with a defined exception allowing explicit parallel detached requests to bypass the guard for concurrent task execution.

Does cross-AI orchestration support handback to non-Anthropic CLIs like Copilot and raw shell?

Cross-AI orchestration handback supports non-Anthropic CLIs including Codex, Copilot, Gemini, and raw shell, enabling spec-kit workflows that require full OpenCode runtime context and CocoIndex semantics.