Back to directory

WrongStack

A free, MIT-licensed AI coding agent, written from scratch as its own kernel rather than a wrapper, with durable project memory, parallel subagents, an inter-agent mailbox and auto-review agents that critique diffs.

WrongStack

WrongStack is a complete, standalone coding agent, with its own compact kernel, provider transports with real server-sent-event streaming, tool executor, permission policy, memory system and multi-agent runtime, rather than a plugin layer or orchestration kit built on top of another CLI tool. It runs across a terminal REPL, a full-screen TUI and a browser UI, all sharing the same permission boundaries.

Project-wide "SAGE" memory persists knowledge across sessions, active Kanban boards support atomic task verification, an inter-agent mailbox links every connected client, and "Chimera" auto-review agents critique diffs. Its provider catalog is pulled live from models.dev. A recent release adds an approval requirement before destructive actions such as disk wipes, infrastructure teardown, dropping container volumes or databases, rsync --delete outside the project, discarding local git work, or publishing to package registries such as PyPI, RubyGems, NuGet or Maven, and redacts secrets from logs and memory.

More in Coding agents

See all 793
  • The City

    github.com

    A native macOS app that shows Claude Code runs as a city, with each project as a building, each team as a floor of robots, and every animation driven by a real event from the claude CLI's stream.

  • rig

    tryrig.ai

    A small, single-binary agent runtime that points at any OpenAI-compatible model, local or hosted, reads and edits a repository's code, runs its tests and remembers what it learned, with a loop implemented in a few hundred lines.

  • Jauvex

    jauvex.reindent.com

    An Electron desktop app that puts Claude Code, Codex and Grok coding-agent sessions side by side and lets you steer them by voice, with a small model handling fast routing decisions.