aicoolies logo
Statewright logo

Statewright

State-machine guardrails for controlling which tools AI coding agents can use at each phase.

Share
open-sourceOpen Source
Visit Website →

Statewright is a guardrail layer for AI coding agents that uses explicit state machines to control what an agent can do at each stage of a workflow. Instead of relying only on prompt instructions, teams can model phases such as plan, implement, test, and review, then constrain tool access for clients like Claude Code, Codex, Cursor, opencode, and related MCP workflows.

Statewright focuses on a specific agent reliability problem: AI coding agents often have too much tool access at the wrong time. Rather than depending only on natural-language instructions, Statewright uses state-machine guardrails so teams can define workflow phases and control what an agent is allowed to do in each phase. That makes it more of an agent-control layer than a coding assistant by itself, and it fits the growing need for explicit policy around autonomous coding workflows.

In practice, the useful pattern is to model phases such as planning, implementation, testing and review, then restrict tools according to the current state. A coding agent might be allowed to inspect files during planning, edit during implementation, run tests in a verification phase and avoid destructive commands unless explicitly permitted. That fits the direction of Claude Code, Codex, Cursor and MCP-style tool use, where the safety problem is less about text generation and more about actions.

Statewright is still a young project, so teams should treat benchmark and reliability claims carefully until they reproduce them in their own repositories. It may require workflow modeling effort before the benefits are obvious, and overly rigid state machines can slow down exploratory development. Even so, the concept is timely: as coding agents become more autonomous, phase-aware tool permissions are a practical way to reduce accidental damage and make agent behavior easier to reason about.

Pricing

Public site and GitHub project are available; commercial terms and long-term licensing should be verified before making procurement claims.

Platforms

State-machine guardrail project for AI coding agents and MCP-oriented workflows, with a public site and GitHub repository.

Categories

Tags

Use Cases

Alternatives

Guardrails AI logo

Guardrails AI

Validate and structure LLM outputs with composable Guards

Guardrails AI is an open-source Python and JavaScript framework for validating and structuring LLM outputs using composable Guards built from a Hub of pre-built validators. It handles structured data extraction with Pydantic models, content safety checks including toxicity, PII detection, competitor mentions, and bias filtering, plus automatic re-prompting when validation fails. The Guardrails Hub offers dozens of validators from regex matching to hallucination detection via LLM judges.

free

NeMo Guardrails

Programmable safety rails for LLM applications

NeMo Guardrails is NVIDIA's open-source toolkit for adding programmable safety rails to LLM applications. It supports five guardrail types — input, dialog, retrieval, execution, and output rails — covering content safety, jailbreak detection, topic control, PII masking, hallucination detection, and fact-checking. The toolkit uses Colang, a domain-specific language for defining conversational constraints, and integrates with OpenAI, Azure, Anthropic, HuggingFace, and LangChain/LangGraph.

free
Claude Code logo

Claude Code

Top Pick

Anthropic's agentic coding CLI

Anthropic's agentic CLI coding tool that delegates complex tasks to Claude directly from the terminal. Understands entire codebases via automatic context gathering, edits multiple files, runs shell commands, and manages Git workflows autonomously. Supports CLAUDE.md for persistent project instructions, integrates with VS Code and JetBrains, and uses Claude Opus/Sonnet with extended thinking for complex architectural decisions. Built for terminal-first developers.

paidOpen Source
Cursor logo

Cursor

The AI-first code editor

AI-first code editor built as a VS Code fork that deeply integrates LLMs into every part of the development workflow. Features Tab autocomplete with multi-line predictions, Cmd+K inline editing, AI chat with full codebase awareness, and Agent mode for autonomous multi-file edits with terminal execution. Supports GPT-4, Claude, and more with automatic context from project files and docs. Includes privacy mode for SOC 2 compliance. The leading AI-native IDE with 100K+ paying users.

freemiumTelemetry

Related Tools

Grok logo

Grok Build

Top Pick

xAI's terminal coding agent with parallel subagents and worktree-aware automation

Grok Build is xAI's terminal-first coding agent for planning, editing, testing, and reviewing code from a local CLI. The early beta exposes subagent controls, worktree mode, headless JSON output, best-of-N parallel attempts, sandbox profiles, and experimental memory. It fits developers comparing Claude Code, Codex, and Gemini CLI for local agentic workflows with deeper parallel execution.

paid
OpenCode logo

OpenCode

Top Pick

Open-source AI coding agent for the terminal

Open-source terminal-based AI coding agent built in Go by the SST team, with a rich TUI (Bubble Tea) supporting 75+ model providers including OpenAI, Anthropic, Gemini, Bedrock, Groq, and OpenRouter. Features vim-like editing, persistent SQLite sessions, and LSP integration for 40+ languages. Fully free with no vendor lock-in, it has rapidly grown to 95k+ GitHub stars.

open-source
Figma Context MCP logo

Figma Context MCP

MCP server that gives coding agents structured Figma context for design-to-code work

Figma Context MCP is an MCP server for giving coding agents structured access to Figma design context during implementation. Instead of copying screenshots or hand-written design specs into prompts, teams can expose layout, component, and context information to agents such as Cursor, Claude Code, and other MCP-compatible coding workflows. It is a strong design-to-code bridge for teams trying to reduce hallucinated UI details and tighten handoff between designers and AI-assisted developers.

open-sourceOpen SourceTelemetry

Grok CLI

Community Grok terminal agent for xAI-powered coding and command-line workflows

Grok CLI is a community command-line interface for using xAI/Grok models from a terminal workflow. It fits developers who want a lightweight, scriptable Grok surface for coding help, command-line experiments, and local agent-style interactions without waiting for a heavier IDE integration. For aicoolies, it belongs in the fast-growing AI CLI agents lane beside Grok Build, Claude Code, Codex, Gemini CLI, and Qwen Code.

open-sourceOpen SourceTelemetry
Webwright logo

Webwright

Microsoft browser agent that turns long-horizon web tasks into reusable Playwright code

Webwright is a Microsoft browser-agent project that asks coding models to write, debug, and reuse Playwright scripts instead of relying on one-off stochastic click loops. The approach gives automation teams a more inspectable artifact: scripts can be logged, reviewed, rerun, and maintained like normal test or scraping code. It is especially relevant for long-horizon browser tasks where teams care about determinism, auditability, and resilience to UI changes.

open-sourceOpen Source

agent-desktop

Accessibility-tree desktop automation engine for deterministic native-app control

agent-desktop is a Rust-native desktop automation engine for AI agents that need structured control of native applications without relying only on screenshots or pixel loops. It exposes accessibility-tree snapshots, stable element references, progressive traversal, and action primitives that can let coding agents and automation stacks operate on Windows, macOS, Electron, and legacy interfaces with lower token cost and better repeatability than pure vision control.

open-sourceOpen Source