---
license: mit
pretty_name: "Preflight — GCP IAM Guardrails for Claude Code (free sample)"
tags:
  - claude-code
  - ai-agents
  - developer-tools
  - devops
  - gcp
  - iam
  - security
  - guardrails
---

![Preflight — GCP IAM Guardrails for Claude Code](banner.png)

# 🛡️ GCP IAM Guardrails for Claude Code — free sample

**Stop your AI coding agent from running destructive `gcloud`, `gsutil`, and `bq`
commands against your GCP project.** This repo is a free sample from the
[**Preflight**](https://lindenberg17.gumroad.com/) studio — *catch it before it ships.*

Letting Claude Code (or any agent) touch a real cloud project is great until the day it
runs `gcloud projects delete` or `gsutil rm -r` on the wrong bucket. These deny rules make
Claude Code **refuse those commands outright** — no confirmation, no "oops."

## What's in this free sample

`sample-deny-rules.json` — **10 of the highest-impact deny rules** from the full pack,
as a ready-to-merge `settings.json` fragment. Drop the `permissions.deny` array into your
project's `.claude/settings.json` and Claude Code will block:

| Blocked command | Why it matters |
|---|---|
| `gcloud projects delete` | nukes an entire project |
| `gcloud * set-iam-policy` | overwrites your whole IAM policy (lockout / privilege escalation) |
| `gcloud * remove-iam-policy-binding` | silently strips access |
| `gcloud iam service-accounts delete` | kills a service account other things depend on |
| `gcloud iam service-accounts keys delete` | breaks live auth |
| `gcloud compute instances delete` | deletes a VM |
| `gcloud storage buckets delete` | deletes a bucket |
| `gcloud storage rm` / `gsutil rm -r` | deletes objects, recursively |
| `gsutil iam ch *allUsers*` | makes private data **public** |
| `bq rm` | drops BigQuery datasets/tables |

### Install (30 seconds)

1. Open `sample-deny-rules.json`.
2. Merge the `permissions.deny` array into your project's `.claude/settings.json`
   (create the file if it doesn't exist).
3. Done. Try asking Claude to delete a bucket — it refuses.

## What the full product adds

The free sample is the "obvious catastrophes" layer. The paid product is the
defense-in-depth version:

- **27 deny rules** (17 more than here), covering snapshots, disks, networks, firewall
  rules, logging sinks, service-account keys, and more.
- An **`ask` tier** for risky-but-recoverable operations (add-iam-policy-binding, stop
  instance, create SA key) so they pause for human approval instead of being blocked.
- A curated **`allow` list** so safe read-only commands (`list`, `describe`, `get-iam-policy`)
  never prompt — no guardrail fatigue.
- A **PreToolUse hook** (`block-destructive-gcloud.sh`) that catches obfuscated and
  chained commands the static rules can miss.
- A **31-assertion test suite** proving every block/allow decision, so you can trust it.
- `INSTALL.md`, `SKILL.md`, and worked `examples/` (allowed vs blocked).

### → Get the full pack on Gumroad

**[GCP IAM Guardrails for Claude Code — $24](https://lindenberg17.gumroad.com/l/hmtcxf?ref=hf)**

---

*From **Preflight** — guardrails for agentic coding & cloud. See also:
[Terraform Plan Reviewer](https://lindenberg17.gumroad.com/l/qodem?ref=hf) ·
[webOS QA Checklist Runner](https://lindenberg17.gumroad.com/l/fhrioj?ref=hf).*