I built guardrails for Claude Code after it told me a fix "worked" when it didn't

Description and use

<p>Just shipped this after getting burned one too many times: <strong>Claude Code Guardrails Pack</strong> — free hooks and skills that stop an AI coding agent from claiming work is done when it never actually checked.</p> <h2> The problem </h2> <p>You've probably hit this if you use Claude Code (or Cursor, Windsurf, any agentic coding tool) for real work:</p> <ul> <li>It tells you a fix works. You ship it. It didn't work — it read the code and decided it looked right, never ran anything.</li> <li>A staged commit almost went out with an API key sitting in the diff.</li> <li>An unattended agent

Problem

<p>Just shipped this after getting burned one too many times: <strong>Claude Code Guardrails Pack</strong> — free hooks and skills that stop an AI coding agent from claiming work is done when it never actually checked.</p> <h2> The problem </h2> <p>You've probably hit this if you use Claude Code (or Cursor, Windsurf, any agentic coding tool) for real work:</p> <ul> <li>It tells you a fix works. You ship it. It didn't work — it read the code and decided it looked right, never ran anything.</li> <li>A staged commit almost went out with an API key sitting in the diff.</li> <li>An unattended agent

Outcome

<p>Just shipped this after getting burned one too many times: <strong>Claude Code Guardrails Pack</strong> — free hooks and skills that stop an AI coding agent from claiming work is done when it never actually checked.</p> <h2> The problem </h2> <p>You've probably hit this if you use Claude Code (or Cursor, Windsurf, any agentic coding tool) for real work:</p> <ul> <li>It tells you a fix works. You ship it. It didn't work — it read the code and decided it looked right, never ran anything.</li> <li>A staged commit almost went out with an API key sitting in the diff.</li> <li>An unattended agent

Evidence

I built guardrails for Claude Code after it told me a fix "worked" when it didn't

Source