Statewright

State machine guardrails that control which tools your AI agent can use in each phase. Define a workflow once, enforce it across Claude Code, Codex, Cursor, opencode, and Pi.

40 tools, zero structure, another debug spiral. Sound familiar? One hook loads a state machine. Your agent gets only the tools allowed in the current phase. No prompt engineering. No jailbreaks. MCP-enforced.

Structure beats reasoning

• Tool enforcement per phase: Destructive tools don't exist in read-only states. The agent can't call what it can't see. Protocol-level enforcement, not prompt suggestions.
• Fewer wasted tokens: Decision checkpoints force progress or fail — no idle looping. Read dedup catches repeated file reads. Edit guards prevent scope explosion.
• Your tools, automatic: The workflow editor discovers your actual toolscape — built-in tools and MCP servers. Build workflows with the tools you already have.