A terminal-native AI coding agent that reads code, edits files, runs commands, and debugs with full human permission control.
WrongStack is easy to set up with trust notes worth reviewing. Check agent compatibility and use-case fit before adding it to your workflow.
Repository setup guidance
gh repo view WrongStack/WrongStack --webOpen the official repository or website.
Check the README for package manager, auth, and platform requirements.
Try it in a small test task inside your agent workflow.
WrongStack is an AI assistant that runs in your terminal. It can read and edit your code, run commands, and help you debug. You stay in control because every action needs your permission. It supports many AI models and works with multiple agents at once.
WrongStack is a terminal-native AI coding agent that reads your code, edits files, runs commands, and reasons through bugs — while you keep your hand on every permission. It drives autonomous goal loops, parallel subagent fan-out, multi-agent Director orchestration, Brain-governed policy decisions, and collaborative debugging. It ships with 37 built-in tools, 17 skills, 7 first-party plugins, and supports ~110 providers from models.dev. Secrets are AES-256-GCM encrypted at rest with a per-machine key; every tool call clears a per-tool permission policy. It offers three interfaces: a plain readline REPL, an Ink/React TUI (--tui), and a standalone web UI. The kernel is only 505 lines, making it highly modular and swappable.
Looks usable, but maintenance, license, or security notes deserve a closer look.
Last commit was about 49 days ago.
65 GitHub stars indicate community interest.
7 open issues signal maintenance load.
MIT license detected.
Automated code generation and refactoring with human oversight
Debugging complex issues by reasoning through code and running tests
Managing multi-file projects with parallel subagent coordination
Spec-driven development: from specification to verified implementation
Learning new codebases by asking the agent to explain and navigate code
Executes arbitrary commands on your machine; always review permissions before approving
Uses third-party AI providers; data may be sent to external servers
Secrets are encrypted at rest but decrypted in memory; ensure your machine is secure
65
Stars
6
Forks
7
Issues
MIT
License
Run large language models locally with a simple CLI. Supports Llama, Mistral, Gemma, and 100+ models with one command.
A public repository of example skills for Claude, demonstrating how to create reusable instructions and resources for specialized tasks.
Anthropic's official CLI agent for coding. Claude-powered terminal agent that reads, writes, and refactors entire codebases.
A terminal-native AI coding agent that reads code, edits files, runs commands, and debugs with full human permission control.
Full-stack developers who want AI assistance with full control, DevOps engineers automating infrastructure-as-code tasks, Technical leads managing large codebases and delegating tasks to AI agents
The setup section provides repository-level starting guidance, not a guarantee of an independently verified installation. Check the official README and release notes.
Executes arbitrary commands on your machine; always review permissions before approving Uses third-party AI providers; data may be sent to external servers
Similar or complementary options to evaluate include Ollama, anthropic-skills, Claude Code.
Jun 8, 2026
3 security/trust notes recorded.
Setup difficulty is 2/5.