opencode

opencode plugin for Claude Code

Use opencode from inside Claude Code for code reviews or to delegate tasks.

Léelo en español.

This plugin is a port of OpenAI's codex-plugin-cc that swaps the Codex backend for the opencode CLI. Each task or review shells out to opencode run --format json, captures the JSONL event stream, and surfaces the resulting session id so jobs can be resumed later with opencode run --continue --session <id>. The same workflow lives on the Claude Code side: slash commands, a rescue subagent, optional stop-time review gate, background job tracking, and result reporting.

What You Get

• /opencode:review — read-only code review of your current git state
• /opencode:adversarial-review — steerable challenge review with focus text
• /opencode:rescue, /opencode:status, /opencode:result, /opencode:cancel — delegate work and manage background jobs
• /opencode:setup — verify the local opencode CLI, ask to install if missing, toggle the stop-time review gate

Requirements

• opencode CLI (opencode --version must work). Install with:
  • npm install -g opencode-ai, or
  • curl -fsSL https://opencode.ai/install | bash (see opencode docs)
• At least one configured provider (Anthropic, OpenAI, OpenRouter, etc.). Run opencode auth login to set one up.
• Node.js 18.18 or later (used by the companion runtime that wraps opencode).
• Claude Code (this plugin runs inside Claude Code's plugin system).

Install

Add the marketplace in Claude Code:

/plugin marketplace add victalejo/opencode-plugin-cc

Install the plugin:

/plugin install opencode@opencode-plugin-cc

Reload plugins:

/reload-plugins

Then run:

/opencode:setup

/opencode:setup tells you whether opencode is installed and at least one provider is configured. If opencode is missing and npm is available, the setup command can install it for you.

If opencode is installed but no provider is configured:

!opencode auth login

After install you should see:

• the slash commands listed below
• the opencode:rescue subagent in /agents

A typical first run:

/opencode:review --background
/opencode:status
/opencode:result

Usage

/opencode:review

Runs a focused opencode code review on the current change set. Useful when you want:

• a review of your current uncommitted changes
• a review of your branch compared to a base ref like main

Use --base <ref> for branch review. Supports --wait and --background. It is not steerable and does not take custom focus text — use /opencode:adversarial-review when you want to challenge a specific decision or risk area.

Examples:

/opencode:review
/opencode:review --base main
/opencode:review --background

This command is read-only and will not perform any changes. When run in the background you can use /opencode:status to check progress and /opencode:cancel to abort it.

/opencode:adversarial-review

Runs a steerable review that questions the chosen implementation and design.

It can be used to pressure-test assumptions, tradeoffs, failure modes, and whether a different approach would have been safer or simpler.

It uses the same review target selection as /opencode:review, including --base <ref> for branch review. It also supports --wait and --background. Unlike /opencode:review, it can take extra focus text after the flags.

Examples:

/opencode:adversarial-review
/opencode:adversarial-review --base main challenge whether this was the right caching and retry design
/opencode:adversarial-review --background look for race conditions and question the chosen approach

This command is read-only. It does not fix code.

/opencode:rescue

Hands a task to opencode through the opencode:rescue subagent.

Use it when you want opencode to:

• investigate a bug
• try a fix
• continue a previous opencode task
• take a faster or cheaper pass with a smaller model

Supports --background, --wait, --resume, --fresh, and --model <provider/model>. If you omit --resume and --fresh, the plugin can offer to continue the latest rescue session for this repo.

Examples:

/opencode:rescue investigate why the tests started failing
/opencode:rescue fix the failing test with the smallest safe patch
/opencode:rescue --resume apply the top fix from the last run
/opencode:rescue --model anthropic/claude-sonnet-4-20250514 investigate the flaky integration test
/opencode:rescue --background investigate the regression