omega-opencode-cli
Use when the user wants to use OpenCode (opencode CLI) for analysis, codebases, or brainstorming from another agent. Uses headless `opencode run` scripts (no MCP). Triggers on "use OpenCode", "ask OpenCode", "analyze with OpenCode", "brainstorm with OpenCode".
来源信息
- 仓库
- oimiragieo/omega-opencode-cli
- 最近来源活动
- 2026年8月13日 23:46
- 检测到的 SKILL.md 语言
- 英语
- 星标
- 0
- Forks
- 0
安装方式
默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。
检查来源文件
决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。
文件资源管理器
15 个文件正在显示 SKILL.md
SKILL.md
来源说明 · 只读预览- name
- omega-opencode-cli
- description
- Use when the user wants to use OpenCode (opencode CLI) for analysis, codebases, or brainstorming from another agent. Uses headless `opencode run` scripts (no MCP). Triggers on "use OpenCode", "ask OpenCode", "analyze with OpenCode", "brainstorm with OpenCode".
- allowed-tools
- Read, Grep, Bash
# Omega OpenCode CLI (headless)
This skill uses the **OpenCode CLI in non-interactive mode** (`opencode run`) so other agents (Claude Code, Cursor, Codex, Copilot, etc.) can run OpenCode from scripts—no MCP server required.
## Overview
Run OpenCode via [scripts/ask-opencode.mjs](./scripts/ask-opencode.mjs). It invokes `opencode run "…" [--model provider/model] [--format json] [--auto]` and returns the response. Requires **Node.js** and the **OpenCode CLI**. For one-time setup, run **/omega-opencode-setup**. See [references/installation.md](references/installation.md) and [references/auth.md](references/auth.md).
## First-time setup
Run **/omega-opencode-setup** or:
```bash
node .claude/skills/omega-opencode-cli/scripts/verify-setup.mjs
```
## How to run OpenCode (headless)
From the project root (where `.claude` lives):
```bash
node .claude/skills/omega-opencode-cli/scripts/ask-opencode.mjs "USER_PROMPT"
```
Options:
- `--model MODEL` / `-m` — `provider/model` (optional `#variant`), e.g. `anthropic/claude-sonnet-5`, `opencode/gpt-5.5`, `openai/gpt-5.2#high`. List with `opencode models`.
- `--json` — maps to `--format json`; script prints extracted assistant text from JSON events.
- `--auto` / `--sandbox` — maps to `opencode run --auto` (auto-approve permissions that are not denied). Prefer for CI.
- `--agent NAME` — OpenCode agent to use.
- `--attach URL` — attach to a running `opencode serve` instance (e.g. `http://localhost:4096`).
- `--timeout-ms N` — wrapper timeout; exit `124` on expiry.
Examples:
```bash
node .claude/skills/omega-opencode-cli/scripts/ask-opencode.mjs "Review @src/main.js for bugs"
node .claude/skills/omega-opencode-cli/scripts/ask-opencode.mjs "Summarize this" --model anthropic/claude-sonnet-5
echo "Explain recursion" | node .claude/skills/omega-opencode-cli/scripts/ask-opencode.mjs
```
## Slash commands
- **/analyze** — headless analysis
- **/omega-opencode** — flexible analysis/brainstorm entry
- **/omega-opencode-setup** — verify Node + OpenCode CLI
## Auth
[references/auth.md](auth.md) — `opencode auth login`, `/connect`, API keys, Zen/Go.
## Failure handling
Do not fabricate OpenCode output. Return the real error and point at verify-setup / auth docs.
## Copy and run
Copy the entire `.claude` folder into the target project. Details: [references/copy-and-run.md](references/copy-and-run.md).
在 GitHub 查看