Rationguard — AI skill for Claude Code

RationGuard

[](LICENSE)

Detect and rebut rationalization patterns in AI agent output.

AI agents make excuses. "Standing by for instructions." "Too complex to fix." "I'll handle it next pass." These sound reasonable but mean the agent stopped working.

rationguard catches these patterns and provides rebuttals — in real-time via [pluk](https://www.npmjs.com/package/@hivecommons/pluk), injected into the agent's system prompt (prevention), or checked against agent output after each response (detection).

---

Prerequisites

• Node.js 20 or newer (engines.node in package.json)
• pluk, installed alongside rationguard below
• tmux, for rationguard attach and for watching live sessions
• The AI CLI you want to supervise (claude, copilot, gemini, goose or codex), already installed and authenticated

`rationguard check` and `rationguard prompt` only need Node.js.

Install

npm install -g @hivecommons/rationguard @hivecommons/pluk

---

Releases

Maintainers publish releases by pushing a `v*` tag. The tag workflow runs the test gate, then publishes `@hivecommons/rationguard` to npm with provenance via [npm Trusted Publishing](https://docs.npmjs.com/trusted-publishers) (GitHub Actions OIDC, `npm` environment) — no long-lived npm token is stored in the repository.

---

Quick Start

One command — creates a tmux session, starts the AI CLI, wires pluk event capture, opens a terminal window for you to interact, and runs rationguard in this terminal to detect rationalizations:

rationguard attach my-agent --cli=claude --rebuttal=send --dangerous

What happens:

1. A tmux session named my-agent is created
2. claude --dangerously-skip-permissions starts inside it (the --dangerous flag maps t