Implement-then-review pattern. Two agents collaborating via pipe chains. Use when you need a second perspective on code changes.
Implement-then-review pattern. Two agents collaborating via pipe chains. Use when you need a second perspective on code changes.
npx skills add https://github.com/joshrotenberg/agent_workshop --skill pairこのコマンドをClaude Codeにコピー＆ペーストしてスキルをインストール
Declarative multi-stage workflows. Define staged pipelines where each stage feeds its result to the next. Use for structured multi-agent tasks.
GitHub integration via gh CLI. List issues, create PRs, comment on issues. Use for agents that bridge GitHub and the Workshop board.
Autonomous board worker. Receives work items, implements them, reports results. Designed for agents that poll the board and execute tasks independently.
Work board driven development. Post typed work items with dependencies, agents poll by role. Use for multi-step features where ordering matters.
Multiple LLM backends in one workshop. Claude and Codex agents working together. Use when different models have different strengths.
Scheduled agent for periodic checks. Use for CI monitoring, PR scanning, health checks, or any polling-based workflow.
| name | pair |
| description | Implement-then-review pattern. Two agents collaborating via pipe chains. Use when you need a second perspective on code changes. |
| license | MIT |
| metadata | {"author":"joshrotenberg","version":"1.0"} |
Two agents: one builds, one reviews. Connected via pipe.
agent(:impl, "You write clean, well-tested code.", max_turns: 15)
agent(:reviewer, "You review code. Do not modify files.",
model: "opus", allowed_tools: ["Read", "Bash"])
ask(:impl, "Implement the caching layer")
|> pipe(:reviewer, "Review for edge cases")
# If reviewer finds issues:
ask(:impl, "Address this feedback: #{result(:reviewer)}")
|> pipe(:reviewer, "Check if the issues are resolved")
If you are the implementor: Write code, include tests, run them before reporting. Be specific about what you changed and why.
If you are the reviewer: Read the diff carefully. Report findings as a prioritized list. Do not modify files. Focus on correctness, edge cases, and style.