Run implementing and reviewing agents with a lead agent
A lead agent takes review findings, has an implementing agent fix them and resubmit, and pings the human only for out-of-scope decisions or repeated failed reviews. ·
Done with a custom-built or unnamed agent
- What they did
- A lead agent coordinates an implementing agent and a reviewing agent. When review findings come back, the lead passes the relevant ones to the implementing agent to fix and resubmit for review. The loop repeats on its own. The commenter is pinged only if a decision exceeds the lead's purview or reviews fail a set number of times.
- How it went
- The source gives no numbers or results. The commenter says they don't have to handle anything unless needed, and can still read what happened in the transcript afterward.
- Worth knowing
- Pick the failed-review limit that triggers a human ping; the source doesn't say what number is used, and it doesn't cover manual testing details.
Try it yourself with Claude Code
In my repo at [path or repo name], act as a lead agent coordinating two subagents: one that implements [feature or list of issues] and one that reviews the work. Have the reviewer report findings, send them back to the implementer to fix and resubmit, and loop until the review passes. Only stop and ask me if a finding needs a decision outside the scope of [the task], or if the same issue fails review [3] times; otherwise finish with a short summary of what changed and the final review result.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: HN · Nov 18, 2026
Related use cases
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Vibe-code and deploy an idle clicker game
Gilbert Pellegrom had his OpenClaw agent build a simple idle game from a conversation, commit it to GitHub and auto-deploy it to Vercel.
Test features on screen and merge fixed PRs
Ryan Carson uses Devin's Test App feature to record and narrate feature clicks, fix bugs, and merge PRs with a !land command.
Code with OpenCode and DeepSeek instead of Claude Code
A developer switched their coding agent from Claude Code to OpenCode with DeepSeek, reporting comparable quality at far lower cost for coding sessions.
Build an app from a simple prompt (partial result)
The poster asked Replit to build an app from a simple prompt, but the login did not work, and they hit the free usage limit while feeding in errors.
Build a trading bot with cursor-agent automation
A commenter is building a personal trading script/bot and uses cursor-agent with composer-1 to automate coding work.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.