Skill

opencode-runtime

Spawns Haiku subagent to run code reviews of diffs/files, implementation plans, technical questions, and analytical tasks via headless OpenCode execution. Shows progress in Claude's agent panel.

developer-tools

automation

npx claudepluginhub apoapps/swarm-code-plugin --plugin swarm-code

Tool Access

This skill uses the workspace's default tool permissions.

Preview

SKILL.md

Similar Skills

claude-code

Use when the user asks to run Claude Code CLI (`claude`, `claude resume`) for review, analysis, implementation, refactoring, debugging, or follow-up specifically through Claude Code.

1 file

guide

codex-delegate

Delegates coding tasks to OpenAI Codex CLI in background via bash execution, retrieving results from temp files. Triggers on Codex/OpenAI mentions for second opinions or parallel runs.

toby-essentials

cache-components

139.4k

Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.

cache-components

Stats

Parent Repo Stars5

Parent Repo Forks0

Last CommitApr 5, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

OpenCode Runtime

Spawn a Haiku subagent that relays your task to OpenCode via bash. Progress shows in Claude's native agent panel.

How to delegate

Agent(
  subagent_type="swarm-code:opencode-worker",
  model="haiku",
  prompt="<your full task with all context inline>"
)

The Haiku worker will:

ACK via SendMessage
Run bash oc-run.sh "<prompt>" → calls opencode run --model <m> "<prompt>"
Return the result via SendMessage

When to use

Code review of a specific diff or file excerpt
Writing an implementation plan with clear requirements
Answering a specific technical question with context
Any analytical task over ~50 tokens where you'd repeat work OpenCode can do

When NOT to use

Tasks that need live file editing (do those yourself)
Short questions you can answer in one sentence
Tasks requiring iterative back-and-forth with the user mid-task

Prompt requirements

Workers CANNOT read files. Include all context inline:

## Task
Review this function for bugs.

## Code
<paste the actual code here>

## Output Format
Bullet list: [SEVERITY] file:line — description. Max 5 findings.

Model setup

Run /swarm-code:init once per project to configure the OpenCode model. The worker automatically uses the configured model — no manual override needed.