Launching today

dynos-work
Self-auditing AI coding agent that learns your codebase
1 follower
Self-auditing AI coding agent that learns your codebase
1 follower
A plugin for Claude Code, Cursor, and Codex. Describe a feature, approve a spec and plan, and it builds in parallel. Then independent agents check it for bugs, security issues, and spec gaps, and fix what they find. Every task adds prevention rules, so your 10th task beats your 1st. Works with permissions ON: allow one binary and you only see the decisions that are yours. Free and open source (MIT).

Free
Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hey Product Hunt 👋 I built dynos-work because AI coding agents kept saying "done" when the work wasn't. In dynos-work, whether a task is complete isn't the model's call. Each task goes through spec → plan → parallel build → independent audit → repair, and a task can't finish while a blocking finding is open. After each task it writes prevention rules from what went wrong, so it gets better at your codebase over time.
It works with permissions ON. You allow one binary, and you only see the decisions that are yours to make. To install in Claude Code:
claude plugin marketplace add dynos-fit/dynos-workWould love feedback, especially on where it gets in your way.